New papal names often drip with meaning. Pope Francis, in 2013, named himself after Saint Francis of Assisi, signifying his dedication to poverty, humility, and peace. Pope Paul VI, in 1963, modeled himself after Paul the Apostle, becoming the first pope to make apostolic journeys to other continents. When Robert Francis Prevost announced on Saturday he would take the name Leo XIV, he gave an unexpected reason for his choice: the rise of AI. The most recent Pope Leo, Prevost explained, served during the Industrial Revolution at the end of the 19th century, and railed against the new machine-driven economic systems turning workers into mere commodities. Now, with AI ushering in a “new industrial revolution,” the “defense of human dignity, justice and labor” is required, he said. With his name choice and speech, Leo XIV firmly marks AI as a defining challenge facing our world today. But also embedded in the name is a potential path forward. Leo XIII, during his papacy, laid out a vision for protecting workers against tech-induced consolidation, including minimum wage laws and trade unions. His ideas soon gained influence and were implemented in government policies around the world. While it's still unclear what specific guidance Leo XIV may issue on artificial intelligence, history suggests the implications of his crusade could be profound. If he mobilizes the world's one billion Catholics against AI's alienating potential as decisively as his namesake confronted industrial exploitation, Silicon Valley may soon face an unexpected and formidable spiritual counterweight.“We have a tradition that views work from a theological perspective. It’s not simply burdensome; it’s where we develop ourselves,” says Joseph Capizzi, dean of theology and religious studies for The Catholic University of America. “Pope Leo XIV is going to be drawing on our tradition to try to make a case for finding work that dignifies human beings—even while making space for AI to do things that human beings will no longer be doing.” Rerum NovarumAt the heart of Leo XIV’s new name choice is Leo XIII’s formal letter Rerum Novarum, which he wrote in 1891. At the time, the Industrial Revolution was upending society. Mechanized production and factory systems generated unprecedented wealth and productivity, but led to the displacement of many agrarian jobs and people to move into overcrowded, unsanitary urban centers in search of work. The jobs there were grueling, unsafe, and paid terribly. The wealth gap widened dramatically, leading to massive social unrest and the rise of communist ideology. In the midst of these many challenges, Leo penned Rerum Novarum, an encyclical that marked the first major example of a pope commenting on social justice. In it, Leo wrote that “a small number of very rich men” had laid “upon the teeming masses of the laboring poor a yoke little better than that of slavery itself.” There now existed as “the gulf between vast wealth and sheer poverty,” he wrote. To combat this trend, Leo explored potential solutions. First, he rejected communism, arguing that workers had a right to the fruits of their own labors. But he also stressed the need for a living wage, time for workers for family and church, and the right to form Christian trade unions. “He was really championing the rights of workers,” says Dr. Richard Finn, director of the Las Casas Institute at Blackfriars, Oxford. In this colorized print from "La Ilustración Española y Americana," Pope Leo XIII directs a phonograph message to the American Catholic people on the occasion of his jubilee, in 1892. (Getty Images—LTL/Heritage Images)These ideas eventually caught hold. One of the first major advocates of minimum wage laws in the U.S. was the priest and economist John A. Ryan, who cited Pope Leo as a significant influence. Many ideas in his text “A Living Wage and Distributive Justice” were later incorporated into the New Deal, when Ryan was an influential supporter of President Franklin D. Roosevelt. In the 1960s, the Catholic Church eventually came out in support of César Chávez and the United Farmworkers (UFW), which Chávez told TIME in 1966 was the “single most important thing that has helped us.” In Australia, Rerum Novarum influenced political leaders who forged a basic wage in that country. And in Mexico, the Rerum Novarum spurred the creation of many Catholic labor unions and mutual aid societies. “It really shaped Catholic activism, with organizations working to ensure that Mexico was neither an unfettered capitalist country nor a Marxist state-owned state,” says Julia Young, a professor at the Catholic University of America. “It was successful in creating Catholic associations that were very politically vocal.” The Church and AIMore than a century after the industrial revolution, a similarly impactful technological revolution is unfolding, amidst many similar economic circumstances. “In terms of similarities between now and then, there was rural to urban immigration changing the workplace, widespread exploitation of workers, and seemingly growing poverty in urban areas,” Young says. “And so you had the church trying to respond to that and saying, ‘We have a different response than Marx or the robber barons.” While Leo XIV hasn’t yet explicitly called for any of the same measures as Leo XIII, it is clear that he believes the rise of AI necessitates some sort of counterweight. And his citing of Rerum Novarum also perhaps reveals a hunger to provoke widespread social change and offer a third path in a two-power arms race. “In our own day, the Church offers to everyone the treasury of her social teaching in response to another industrial revolution,” he said on Saturday. Across the world, people are expressing intense anxiety about AI causing job displacement. (Some economists contend that these fears are overblown, however.) Like in the industrial revolution, the initial spoils of AI are flowing to a few ultra-powerful companies. And AI companies have also reinforced some of the worst aspects of predatory global capitalism systems: OpenAI, for instance, outsourced some of its most grueling AI training to Kenyan laborers earning less than $2 an hour. Leo’s interest in this area continues that of Pope Francis, who became increasingly vocal about the threats to humanity posed by AI in his later years. Last summer at the G7 Summit, he called for an international treaty to regulate AI, arguing that it could exacerbate social tensions, reinforce dominant cultures, and undermine education. “We would condemn humanity to a future without hope if we took away people’s ability to make decisions about themselves and their lives, by dooming them to depend on the choices of machines,” he said. Some leaders have signaled the importance of prioritizing workers’ rights during the AI revolution, like Senator Josh Hawley. But until a coherent political movement emerges, moral leadership on human dignity in the face of AI may flow from the church, and Pope Leo’s outspoken leadership. “He’s saying AI is going to change the workplace—but it's got to change it in a way that fits with the dignity of employees,” says Dr. Finn. 

Apple users—specifically those who use Siri through products such as Macbooks, iPhones, and Apple TVs—may be entitled to make a claim after Apple’s class action lawsuit settlement, worth $95 million dollars, regarding the voice-activated assistant.The settlement comes from a lawsuit filed in 2021 by Californian Fumiko Lopez, who claimed that Apple, via Siri, conducted “unlawful and intentional interception and recording of individuals’ confidential communications without their consent and subsequent unauthorized disclosure of those communications.”“Apple intentionally, willfully, and knowingly violated consumers’ privacy rights, including within the sanctity of consumers’ own homes where they have the greatest expectation of privacy,” the lawsuit stated. “Plaintiffs and Class Members would not have bought their Siri Devices, or would have paid less for them, if they had known Apple was intercepting, recording, disclosing, and otherwise misusing their conversations without consent or authorization.”In 2019, Apple published a statement titled "Improving Siri’s privacy protections," in which they said they hadn't "been fully living up" to their "high ideals" and vowed to issue improvements.Apple agreed to the settlement on Dec. 31, 2024. According to the settlement website: "Apple denies all of the allegations made in the lawsuit and denies that [they] did anything improper or unlawful."The website also provides information about who is eligible to file a claim and the deadlines they need to adhere to. Here’s what you need to know about how you can file a claim:Who is eligible to file a claim?People eligible to make a claim include those who owned or purchased a Siri device—which includes the iPhone, iPad, Apple Watch, MacBook, iMac, HomePod, iPod touch, Apple TV—between Sept. 17, 2024 and Dec. 31, 2024. They must have “purchased or owned a Siri Device in the United States or its territories and enabled Siri on that device.”According to the settlement agreement, eligible parties also should have “experienced an unintended Siri activation during a confidential or private communication.”Those not eligible include Apple employees, legal representatives, and judicial officers assigned to the case.How can you make a claim and when is the deadline?Claimants can submit a claim form via the settlement website, and can submit claims for up to five Siri devices. The deadline to make a claim is July 2, 2025. This is also the deadline to opt out of the payment, which would allow the customer to keep their right to bring any other claim against Apple arising out of, or related to, the claims in the case. Some of those eligible to make a claim may have received a postcard or an email—with the subject line “Lopez Voice Assistant Class Action Settlement”—notifying them about the settlement. This correspondence would likely include a Claim Identification Code and a Confirmation Code. Per the settlement website, people can use these codes when making a claim, but eligible Apple customers who haven’t received any correspondence can still file a claim.When can you expect to receive payment?On August 1, 2025, the courts are due to host a final approval hearing, but there could still be appeals. Payments will only be issued after any appeals are resolved. The settlement website is set to keep customers updated on timings and payment schedules, as and when that information is available.

This week, President Trump traveled to the Middle East on a business and diplomacy mission, during which he greenlit the sale of hundreds of thousands of American-made AI chips to firms in the UAE and Saudi Arabia. These deals signal a major shift in the U.S.’s approach to cutting-edge AI technology. Previously, U.S. leaders had focused on limiting access to ultra-powerful chips, especially to countries that might pose national security threats. Now, Trump is using them as leverage for his larger trade ambitions. While Trump was at the Saudi-U.S. Investment Forum in Riyadh—held in parallel with his visit—the White House announced that Saudi Arabia was committing $600 billion in investments in the United States, “building economic ties that will endure for generations to come,” Onstage at the conference, Nvidia CEO Jensen Huang announced his company was entering a massive partnership with Humain, a new company owned by the Saudi kingdom’s Public Investment Fund, and sending them hundreds of thousands of chips. Rival chipmaker AMD announced its own $10 billion Saudi Arabian project. Another deal in the works could send hundreds of thousands of chips to the Emirati firm G42. While Trump allies heralded the deal as mutually beneficial to all parties, some national security experts have concerns about the longterm impacts of spreading these chips around the world. “AI chips should not be bargaining chips for broader trade deals,” says Janet Egan, a senior fellow at the Center for a New American Security (CNAS). “They underpin US AI dominance, and we have to be really careful to not make short term decisions that might be beneficial for trade in the near term, but cede AI leadership in the longer term.”Early this year the Chinese company Deepseek revealed that it had developed a very powerful model mostly using Nvidia chips obtained before the Biden administration closed an export loophole in 2023, heightening the intensity of the race. President Biden, in his final weeks, ratcheted up export controls, including limits on countries in the Gulf. Last week, the Trump administration ripped up those rules, with a spokesperson calling them “overly complex, bureaucratic” and saying they “would stymie American innovation.” They then switched to a new tack: linking countries’ access to AI chips with larger trade negotiations. Transitioning to a negotiation-based approach, the administration argued, could allow for more flexibility from country-to-country and allow Trump to secure key business concessions from Middle Eastern partners. Business and governments in the Middle East have massive ambitions for AI, aiming to position themselves at the forefront of this emerging technology. They benefit from several strategic advantages to do so, including access to boundless energy, free-flowing capital thanks to oil and sovereign wealth funds, and a lack of government restrictions—allowing them to rapidly push through massive infrastructure projects. But until now, the Middle East had lacked one crucial puzzle piece: Access to cutting-edge American chips from companies like Nvidia.Now, the amount of chips that U.S. companies will reportedly send to the UAE and Saudi Arabia is massive: “We're talking about something larger than any AI training system that exists in the world today," says Alasdair Phillips-Robins, a fellow at the Carnegie Endowment for International Peace. Conceivably, the ultra-powerful models built with this training system could synthesize automated cyber-attacks, intelligence collection, and weapons development. This is potentially problematic to some U.S. analysts, given Saudi Arabia and the UAE’s close ties with China. In previous years, American spy agencies issued warnings that G42 could be a conduit for siphoning advanced American technology to China. G42 denied any connections to the Chinese government or military.“If you think about which country should be leading the future of potentially the most critical and transformative technology we've ever had, I would not want that to be a non-democratic authoritarian regime,” Egan says. On Tuesday, the House Select Committee on the Chinese Communist Party, led by Michigan Republican John Moolenaar, wrote on Twitter that the new chip deals “present a vulnerability for the CCP to exploit.” Sam Winter-Levy, another Carnegie Endowment fellow, worries that the deal will encourage U.S. AI companies to move to the Gulf, where they might get better deals on energy and avoid U.S. regulations and community pushback. Prominent U.S. companies have wasted no time in seizing the new opportunity presented by the Trump administration. OpenAI’s Sam Altman, Nvidia’s Jensen Huang, and AMD’s Lisa Su all attended the Saudi-U.S.Investment Forum. The AI startup Scale AI—which has a partnership with the U.S. government to develop AI safety standards—announced its intentions to open an office in Saudi Arabia. Google, too, advanced an AI hub in the country. "You could end up in a position where some large proportion of U.S. computing power has been offshored to a bunch of states that can wield leverage over U.S. foreign policy to shape it in ways that may not align with US national interests,” says Winter-Levy,. Winter-Levy also contends that these AI chip deals go against Trump’s past emphasis on an “America first” foreign policy approach. "This is offshoring data centers that could be built in the United States. This is offshoring chips that could be going to US tech companies,” he says. “It's hard to reconcile this with an America First approach to industrial policy or economic policy in general.”

Just a few months ago, the crypto industry seemed unstoppable in Washington. It had the support of a pro-crypto president in Donald Trump, a slew of new pro-crypto legislators in both parties, and newly elevated regulators who pledged to not impede the industry’s growth. Many assumed the speedy passage of pro-crypto legislation as a foregone conclusion after Trump asked Congress to send him a stablecoin bill to sign by August. That momentum hit a major snag over the last few days, as Trump’s expanding investment in the industry coincides with a revolt from Democrats who had previously supported the leading crypto legislation. On Tuesday, California Rep. Maxine Waters, the ranking Democrat on theshot into the top ten stablecoins by market capitalization. In the Senate, a group of nine Democrats announced they would not support a stablecoin bill, called the GENIUS Act, without major changes, significantly narrowing its pathway to 60 votes. Meanwhile, Senate Banking Committee staff and Massachusetts Sen. Elizabeth Warren, the ranking Democrat on the committee, circulated a memo to her fellow Senate Democrats urging them to demand amendments that might address the bill’s national security concerns. “If Congress is going to supercharge the use of stablecoins and other cryptocurrencies, it must include safeguards that make it harder for criminals, terrorists, and foreign adversaries to exploit the financial system and put our national security at risk,” read the memo, which was obtained by TIME. The GENIUS Act is still headed for a vote in the Senatesaid he was open to making changes to reach a compromise that addresses Democrats’ concerns. Here are some of the major objections to the bill, and how the fight may play out this week. Conflict of interest concernsStablecoins are cryptocurrencies designed to hold the value of a U.S. dollar. For many lawmakers on both sides of the aisle, passing a stablecoin bill seemed more feasible this year than tackling a larger crypto market structure bill, especially because stablecoins are less volatile and their value is usually tied to actual money sitting in a bank. Read More: What Are Stablecoins?But in March, Trump’s World Liberty Financial announced a new stablecoin, leading to concerns that the new legislation would essentially give Trump even moreexecutive order placing independent financial regulators like the FTC, FCC and SEC under his own control.)Trump has only escalated his crypto dealings. Last week, World Liberty Financial announced that an Emirati company planned to use the firm’s new stablecoin for a $2 billion investment in Binance, the world’s largest cryptocurrency exchange. Trump also announced that he would host an exclusive dinner for top investors of his $TRUMP meme coin—which Republican Senator Cynthia Lummis of Wyoming, aadmitted “gave [her] pause.”Waters had been working on stablecoin legislation for years. But last month, she reversed course, saying that she opposed any bill that would allow Trump to own a stablecoin. On Tuesday, she walked out of a joint House hearing on crypto, laterWaters then staged her own hearing on stablecoins. Notably, however, several Democrats remained at the original hearing, including Rep. Stephen Lynch of Massachusetts, the ranking member on a subcommittee focused on digital assets, and Rep. Angie Craig of Minnesota. "This is a really important conversation. I'm here because I think we need to be engaged, and part of the discussion," Craig said. Craig, however, agreed that Waters was raising important issues. “It's important and it's legitimate to call out the self-dealing from the Trump administration related to hawking meme coins from the White House,” she said. “It's corrupt, it's wrong, and it makes this process of coming together to regulate crypto more partisan."National Security ConcernsWhile some Democrats are focused on stopping Trump from owning a stablecoin while he’s in office, others are concerned that the current stablecoin bills in Congress could have unintended consequences. Warren, who has long been a crypto skeptic, has particularly honed in on the ripple effects on national security, arguing that the bill would make it easier for terrorists and malicious state actors to steal and cash out illicit funds. In February, hackers backed by the North Korean government stole $1.5 billion in cryptocurrencies from the crypto exchange Bybit, as part of a larger continuing effort to steal crypto funds from around the world. The Bybit hack was the largest in crypto history—and foreign policy experts believe that the stolen funds are being used to fund the development of missile and nuclear weapons technology. So Warren and Banking Committee staffers circulated a memo on Monday, which calls for changes to the GENIUS Act, including the implementation of strict anti-money laundering requirements on exchanges handling digital assets. It argues that the bill should extend U.S. sanctions laws to stablecoins, and that stablecoin issuers should be required to monitor blockchains and report criminal activity. The nine Democrats who revoked their support of the GENIUS Act now hold significant leverage over the bill. It is not clear what changes to the bill would be enough to regain their support. “We've been very clear to our Republican colleagues for weeks about the changes that we need,” Virginia Sen. Mark Warner, one of those nine Democrats, told TIME on Tuesday. Arizona Sen. Ruben Gallego, who led the Democrats’ statement opposing the bill, told TIME that his priorities were beefing up consumer protections and national security issues. “We can tighten up the ‘who can issue, what country can issue’ question,” he says. “It’s incredibly important when it comes to closing some of the Tether loopholes.”Democratic Sen. Angela Alsobrooks of Maryland, a co-sponsor of the bill, told TIME she believes that the bill should require crypto companies dealing with stablecoins to adopt anti-money laundering (AML) and countering the financing of terrorism (CFL) rules. “We still have a little time, but everybody's motivated, and we're all working together to try to get to the best place we can,” she says. “We want to make sure that all of the concerns around national security are addressed.”Republican Sen. Bill Hagerty of Tennessee, one of the bill’s authors, appeared unfazed by the challenges. “I’m beyond optimistic. I’m confident it will pass,” he told TIME.  Independent Vermont Sen. Bernie Sanders announced that he would host a livestream with other critics of the GENIUS Act on Wednesday to discuss how it “threatens the stability of our financial system.” The crypto industry is continuing to push for the bill’s passage. Dante Disparte, a leader at the stablecoin issuer Circle, tells TIME that more harms come from the absence of legislation. "Past failures to pass bipartisan stablecoin legislation have harmed U.S. consumers, markets, national security, and dollar competitiveness,” he wrote in an email, citing the failure of the foreign stablecoin project Terra-Luna in 2022. 

As AI enters the workforce and seeps into all facets of our lives at unprecedented speed, we’re told by leaders across industries that if you’re not using it, you’re falling behind. Yet when AI’s use in art enters the conversation, some retreat in discomfort, shunning it as an affront to the very essence of art. This ongoing debate continues to create disruptions among artists. AI is fundamentally changing the creative process, and its purpose, significance, and influence are subjective to one’s own values—making its trajectory hard to predict, and even harder to confront.Miami-based Panamanian photographer Dahlia Dreszer stands out as an optimist and believer in AI’s powers. She likens AI’s use in art to the act of painting or drawing—simply another medium that can unlock creative potential and an artistic vision that may have never been realized without it. Using generative AI models like Stable Diffusion, 3.5, Midjourney, Adobe, Firefly, and Nova, Dreszer trained an AI image generator on her style for over a year, instructing it to produce artwork with her sensibilities, with one piece in her current exhibition produced entirely by AI. Entitled “Bringing the Outside In,” Dreszer calls the show a “living organism.” (It is on display until May 17, 2025 at Green Space Miami.) Her vivid, maximalist still lifes depict layered familial heirlooms, Judaica, flowers, and textiles made by Panamanian indigenous women. Attendees can interact with an AI image generator in the exhibition to produce their own artworks in Dreszer’s style, telling the machine in a sentence or two what they want it to produce, and in seconds, an artwork is created. Also as part of the show, Dreszer programmed an AI-generated clone of herself, which looks and speaks like her, to guide visitors via video chat through the space. This interview has been lightly edited for length and clarity. An AI-collaborated piece in Dreszer's "Bringing the Outside In" exhibition. Courtesy Dahlia Dreszer TIME: Take me back to the first moment you realized AI could enhance your art. What about AI drew you in? What did you feel? Dreszer: I believe technology is here to supercharge us. When generative AI entered the mainstream, I knew I wanted to get my hands dirty right away. I was already in the world of NFTs, but this was a different conversation. It took over a year of experimentation and dialogue with image generators to feel comfortable finally creating a piece to include in a body of work. This exhibition includes one piece I made in collaboration with AI. I personalized an AI image model on what the exhibition means, feels like, and looks like, feeding it images embodying my style. I included the Florida Everglades in the foreground, reflecting the landscape where I'm living today. I’m not only interested in AI and art, but also in adding nature to that conversation. I’ve hung flowers on top of this piece that fall onto the frame or the ground when they die, allowing nature to do its thing. I have not intervened physically. I believe nature, art and technology can coexist nicely.I actually thought that all the pieces in your exhibition were produced by AI.That's also the intention, right, because they are not. I'm always trying to play with the viewers, to disorient, because everything is not what you see at first glance. There's no artificial enhancements in most of these works, but just the fact that you think there are—I find that narrative interesting. What inspired you to create a clone for this exhibition? My clone is so fun. I'm trying to pose questions to the community as they engage with these works: Moving forward, what does it mean for relationships when we're speaking to a machine as if it was a human, and we cannot know the difference? What is our role as humans if we have clones that can mimic what we do? I want to see how that dialogue evolves. There's a practicality as well. The clone guides you through the show, probably better than I can. It's trained on what I know, but as a machine, it's supercharged. Objects seen in Dreszer's photos are brought to life in her physical exhibition space. Courtesy Dahlia DreszerWhy did you include your clone?I wanted to have an AI version of myself to guide viewers and answer questions, to educate others in order to demystify AI. Through the clone, I can humanize the technology, “the art of the possible,” of incorporating technology into artistic workflows. Will you keep your clone after the exhibition? Will you educate it about other parts of yourself? I'm very interested in continuing the relationship with her. I'm working through ideas and ways to train her. I haven't shared it yet, but there are different personas of the clone. I'll be fine tuning and creating different versions based on the relationship I want her to have with the audience she's engaging with.Some critics would call the use of AI in art “cheating.” What do you say to those critics?I’d love to have a conversation to understand how that opinion was formed. I’d encourage them to see it as a collaboration. Many people don’t understand the process and the time it takes. I would invite critics to dive deeper, and think about it not just as: “I put in a prompt, it makes art, then I'm done.” It's a long process.But this relationship between technology and the arts is not new. We’ve had disruptions in art through technology before. This is just more aggressive, intrusive, and rapid in its speed and pace of innovation. Two of Dreszer's works featured at the exhibition. Courtesy Dahlia Dreszer What specific challenges have you faced so far using AI in your art?Oftentimes the outputs are not what I wanted. As an artist, I have high expectations. I like to control the visualization so it’s highly stylized, curated, and composed. With AI, that control goes away, because AI has its own intelligence and creativity, no matter how good the prompt is. It's a hard and frustrating yet also enlightening process; it may not create what you wanted, but it can make something you didn't know you wanted. Then there's technical things it doesn't know how to do, but eventually will. It’s not great with certain renders or visualizations. What scares and excites you about where AI is headed for the next generation of artists?I'm mostly excited because of the rapid pace. Updates to generative AI software happen in a matter of weeks. There’s also a healthy competition in the market, which means that as users, our needs are being satisfied quicker than ever. Our feedback is being incorporated and the tools are changing. You asked about fears. AI is entering our workflows and industries in one way or another. Will we accept it? Deny it? Who will fall behind, and who will be at the forefront? I’m more excited than fearful, but I see why others may be fearful. It disrupts our workflows, and if we're not ready to change or learn new skills, it can be scary. Will collaboration with AI replace collaboration between artists?No, no, no. There are many examples of how me and several artists have collaborated with AI. One artist came to me with her artistic vision and her words, and I used my prompt engineering skills and knowledge of AI systems, and together, we created an AI piece that was her vision come to life—this beautiful red textile tree that had a huge trunk.An AI-collaborated piece made together with artist Karla Kantorovich. Courtesy Dahlia Dreszer and Karla KantorovichAs an artist, there is a journey one goes through when creating. When you use AI, does it still allow you to access this other-worldly experience of the creative process?There's definitely parts of the creative process that AI is not inclusive of. So for example, when I'm making AI art, I'm not painting, or getting my hands dirty. There's physicalities that are not included in that journey. But I think that's similar to any medium. So let's say I'm choosing to use my camera as my tool and not a paint brush. There's also experiences that are missed out through my photographic artistic process, that if I were using a paintbrush or another tool, would be a different journey. So that’s why I see generative AI art as its own medium, and each medium comes with its own journeys and processes that are exclusive to that medium, right?Do you see the term “post-human” as an accurate way to reflect this era we are entering in art?I would divert a bit from “post-human.” I see AI more as a booster, not a replacer, but an accelerator and an enabler. So, if “post-human” means it's a replacement, then I would lean in more to the perspective of AI as a turbo supercharger that us humans can carry with us to bolt forward. I think it could replace mundane tasks that we may not want to do. And that's where the beauty of the collaboration comes in, where we give it these tasks so our human brains reach our fullest potential, because then the low value tasks we can outsource into generative AI.How do you think historians will look back on this particular era of rapid expansion with AI?We are in the foundation era. Everyone knows what ChatGPT is. We've passed the point of inflection, and now we're at a point where industries, individuals, businesses, and creatives are finding their place in AI. How are we adapting–or not–to it? Time is of the essence. What we decide to do now, literally today, versus in a week or two, or three, or in a month, will define the next five to 10 years. 

Shortly after my mother’s 89th birthday I called her landline via Skype. It was our last such communication ever. Don’t worry, mom’s fine. Skype, however, is as dead as the dire wolf. Deader, probably, because nobody is trying to figure out how to bring Skype back. My mother was born before humans discovered how to use antibiotics and she’s still going. Skype was born in 2003 and only just made it past 20 years. Technological change is inevitable, but given the current manic pace of innovation, we are more accustomed to experiencing this as the arrival of shiny new tools, not the departure of useful (if slightly simplistic) old ones. When my grandmother was born, there were no planes. When my mother was born, there were no transistors. When I was born, there were no mobile phones. Those technologies are still going strong. In 2019 Skype was declared one of the top 10 most downloaded apps of the 2010s, above TikTok and YouTube and Twitter. That’s only six years ago. I haven’t even been to the gynecologist since then. Is there a word for the sense of loss you experience when you outlive a technology that changed your life? I know some people feel enough nostalgia for BlackBerrys and Sony Walkmans and even horse gas masks that they have become collectible, but when software goes, what remains? How do we memorialize and mourn a series of zeros and ones that opened a whole new world to us? I was just old enough when Skype came on the horizon to really appreciate it. As a traveler and an expatriate, I made a lot of long-distance phone calls. A way of reaching lovers, family, and friends when you needed to, long-distance calls had their own kind of romance, especially if you enjoyed having a conversation where every sentence you uttered cost you—and sometimes the person you were talking to—about two dollars. It made you measure your words. If your father was particularly frugal, as mine was, you’d never even try to sing all of “Happy Birthday to You,” for example, for fear of ruining his whole year. In fact, what a long-distance call often consisted of was long pauses as people tried to think of things to say that were worth the money. In my family, we couldn’t conjure conversation that rich fast enough, so we’d exchange pleasantries, hang up, and then curse ourselves for wasting money on such a nothing call. Friends of mine used to take notes before they called for maximum efficiency. In some countries I visited, you had to pay for a certain number of minutes first, hand over the number you wanted to call, and then go sit in a booth and wait to be connected. The pressure to fill those prepaid minutes with worthwhile content was intense. Skype was not the only solution to this. There were, briefly, specialized international calling companies where you could pick one or two countries and call them for a bargain rate of, say, 20 cents a minute. (Much of my phone conversation with my father during that era was spent marveling with him at how cheap it was.) But Skype was one of the earliest and easiest to use, and it called landlines for a few cents, so if you could not pry your beloved elders’ hands away from their handsets, it was a godsend. Small talk was possible! You could digress! You could sing all of “Happy Birthday to You” and get halfway through “For He’s a Jolly Good Fellow” before you realized that you actually didn’t miss singing as much as you thought you did. Invented in 2003 by some now-billionaire northern Europeans, Skype, which used the internet rather than phone lines to connect people, was sold off to eBay in 2005, and eventually ended up at Microsoft, which is retiring it in favor of Teams. As technology goes, this is a familiar cycle: innovation, monetization, ruination. Skype is like that alt-rock group whose live concert was the first you ever saw, but who kept switching record labels and eventually disbanded. At least with a concert tour, you get a T-shirt. All we Skypers have is a vestigial blue S bubble on our phones. Perhaps Skype’s appeal to the less technologically savvy was what doomed it. I only ever used Skype for one thing: to call my mother’s landline. I didn’t use it for messaging or video. It offered translation and payments and redesigns, all of which I ignored. I bristled when it briefly started sending me daily news items. I acknowledge my complicity in its demise. For me, Skype was like the BOOST button on my mother’s telephone, which turns the volume up; it had a limited but crucial utility.Now Skype is gone. Though each of her descendants has tried to get her to use any of the communication methods invented after 1876, my mother still wants to pick up the receiver of a ringing phone, like she always has. For her, zooming is what cars do and FaceTiming is what folks used to call coming over for a cuppa. I will now call her (for free) through one of the other apps, which is only slightly more complicated and allows her to keep her feet planted in the technological era in which she feels safe. But it feels like the distance is getting wider, that the rubber cord between us is reaching the outer limit of its stretchiness. As digital communication grows more sophisticated, she seems older, farther away, less reachable. I can see and hear everybody else clearly, but mom is just a whisper. And I can’t help worrying that it’s not just inventions that cannot keep up that get abandoned sooner—it’s people. I know Skype was just a stage, and pouting over its demise is like wishing cocoons never became butterflies, but still, I would have liked a T-shirt.

On April 28, the House of Representatives passed the first major law tackling AI-induced harm: the Take It Down Act. The bipartisan bill, which also passed the Senate and which President Trump is expected to sign, criminalizes non-consensual deepfake porn and requires platforms to take down such material within 48 hours of being served notice. The bill aims to stop the scourge of AI-created illicit imagery that has exploded in the last few years along with the rapid improvement of AI tools. While some civil society groups have raised concerns about the bill, it has received wide support from leaders on both sides of the aisle, from the conservative think tank American Principles Project to the progressive nonprofit Public Citizen. It passed both chambers easily, clearing the House with an overwhelming 409-2 vote. To some advocates, the bill is a textbook example of how Congress should work: of lawmakers fielding concerns from impacted constituents, then coming together in an attempt to reduce further harm. "This victory belongs first and foremost to the heroic survivors who shared their stories and the advocates who never gave up," Senator Ted Cruz, who spearheaded the bill in the Senate, wrote in a statement to TIME. "By requiring social media companies to take down this abusive content quickly, we are sparing victims from repeated trauma and holding predators accountable."Here’s what the bill aims to achieve, and how it crossed many hurdles en route to becoming law.Victimized teensThe Take It Down Act was borne out of the suffering—and then activism—of a handful of teenagers. In October 2023, 14-year-old Elliston Berry of Texas and 15-year-old Francesca Mani of New Jersey each learned that classmates had used AI software to fabricate nude images of them and female classmates. The tools that had been used to humiliate them were relatively new: products of the generative AI boom in which virtually any image could be created with the click of a button. Pornographic and sometimes violent deepfake images of Taylor Swift and others soon spread across the internet. When Berry and Mani each sought to remove the images and seek punishment for those that had created them, they found that both social media platforms and their school boards reacted with silence or indifference. “They just didn’t know what to do: they were like, this is all new territory,” says Berry’s mother, Anna Berry. Anna Berry then reached out to Senator Ted Cruz’s office, which took up the cause and drafted legislation that became the Take It Down Act. Cruz, who has two teenage daughters, threw his political muscle behind the bill, including organizing a Senate field hearing with testimony from both Elliston Berry and Mani in Texas. Mani, who had spoken out about her experiences in New Jersey before connecting with Cruz’s office during its national push for legislation, says that Cruz spoke with her several times directly—and personally put in a call to a Snapchat executive asking them to remove her deepfakes from the platform. Mani and Berry both spent hours talking with congressional offices and news outlets to spread awareness. Bipartisan support soon spread, including the sign-on of Democratic co-sponsors like Amy Klobuchar and Richard Blumenthal. Representatives Maria Salazar and Madeleine Dean led the House version of the bill. Read More: Time 100 AI 2024: Francesca ManiPolitical wranglingVery few lawmakers disagreed with implementing protections around AI-created deepfake nudes. But translating that into law proved much harder, especially in a divided, contentious Congress. In December, lawmakers tried to slip the Take It Down Act into a bipartisan spending deal. But the larger deal was killed after Elon Musk and Donald Trump urged lawmakers to reject it. In the Biden era, it seemed that the piece of deepfake legislation that stood the best chance of passing was the DEFIANCE Act, led by Democrats Dick Durbin and Alexandria Ocasio-Cortez. In January, however, Cruz was promoted to become the chair of the Senate Commerce Committee, giving him a major position of power to set agendas. His office rallied the support for Take it Down from a slew of different public interest groups. They also helped persuade tech companies to support the bill, which worked: Snapchat and Meta got behind it. “Cruz put an unbelievable amount of muscle into this bill,” says Sunny Gandhi, vice president of political affairs at Encode, an AI-focused advocacy group that supported the bill. “They spent a lot of effort wrangling a lot of the companies to make sure that they wouldn't be opposed, and getting leadership interested.”Gandhi says that one of the key reasons why tech companies supported the bill was because it did not involve Section 230 of the Communications Act, an endlessly-debated law that protects platforms from civil liability for what is posted on them. The Take It Down Act, instead, draws its enforcement power from the “deceptive and unfair trade practices” mandate of the Federal Trade Commission. “With anything involving Section 230, there's a worry on the tech company side that you are slowly going to chip away at their protections,” Gandhi says. “Going through the FTC instead was a very novel approach that I think a lot of companies were okay with.”The Senate version of the Take It Down Act passed unanimously in February. A few weeks later, Melania Trump threw her weight behind the bill, staging a press conference in D.C., with Berry, Mani, and other deepfake victims, marking Trump’s first solo public appearance since she resumed the role of First Lady. The campaign fit in with her main initiative from the first Trump administration: “Be Best,” which included a focus on online safety. A Cruz spokesperson told TIME that Trump’s support was crucial towards the bill getting expedited in the House. “The biggest challenge with a lot of these bills is trying to secure priority and floor time,” they said. “It’s essential to have a push to focus priorities—and it happened quickly because of her.”"Today's bipartisan passage of the Take It Down Act is a powerful statement that we stand united in protecting the dignity, privacy, and safety of our children," Melania Trump said Monday. "I am thankful to the Members of Congress — both in the House and Senate — who voted to protect the well-being of our youth."Support is broad, but concerns persistWhile the bill passed both chambers easily and with bipartisan support, it weathered plenty of criticism on the way. Critics say that the bill is sloppily written, and that bad faith actors could flag almost anything as nonconsensual illicit imagery in order to get it scrubbed from the internet. They also say that Donald Trump could use it as a weapon, leaning on his power over the FTC to threaten critics. In February, 12 organizations including the Center for Democracy & Technology penned a letter to the Senate warning that the bill could lead to the “suppression of lawful speech.” Critics question the bill’s effectiveness especially because it puts the FTC in charge of enforcement—and the federal agency has been severely weakened by the Trump administration. At a House markup in April, Democrats warned that a weakened FTC could struggle to keep up with take-down requests, rendering the bill toothless. Regardless, Gandhi hopes that Congress will build upon Take It Down to create more safeguards for children online. The House Energy and Commerce Committee recently held a hearing on the subject, signaling increased interest. “There's a giant movement in Congress and at the state level around kids' safety that is only picking up momentum,” Gandhi says. “People don't want this to be the next big harm that we wait five or 10 years before we do something about it.”For Mani and Berry, the passage of Take It Down represents a major political, legal, and emotional victory. “For those of us who've been hurt, it's a chance to take back our dignity,” Mani says.

For the second time, TIME is awarding the World's Top EdTech Companies 2025, in partnership with Statista, a leading international provider of market and consumer data and rankings. The result of this quantitative study: 350 companies dedicated to the development and provision of educational technologies, encompassing both products and services. Here's how the winners were selected.MethodologyThe research project "World's Top EdTech Companies 2025" is a comprehensive analysis conducted to identify the top performing EdTech companies in the United States. First, companies had to primarily focus on developing and providing educational technologies, products, or services. The 350 companies with the highest score were awarded as "World's Top EdTech Companies 2025."The ranking is built on two pillars: financial strength and industry impact. Statista gathered and scrutinized data from over 7,000 companies through desk research, online application forms, and collaborations with other data and market intelligence companies. A company received scores in each of these dimensions, which were then combined into an overall score.For the first dimension, financial strength, Statista analyzed revenue, employee, and funding data, obtained from publicly available sources like annual reports, company websites, through media monitoring, and via databases. Additionally, company disclosures submitted via an online application form, which was freely accessible via the TIME website, were considered.For the evaluation of the industry impact, Statista cooperated with HolonIQ and LexisNexis® Intellectual Property Solutions to assess companies in different impact dimensions, encompassing factors such as:quality and impact of their product/service portfolioquantity and value of a company’s IP (intellectual property) portfolioIn addition, Statista researched various metrics around web traffic, to assess the relevance of the products and services to the respective target groups.The evaluation process involved assessing all relevant companies within the global EdTech sector that exhibited above-average performance on financial metrics in their geographical region. Subsequently, the selection was narrowed down to include only those companies that not only excelled financially but also demonstrated significant industry impact within their specific categories. This selection process allowed us to establish our final ranking of the World's Top EdTech Companies 2025. Once the data was collected and evaluated, it was consolidated and weighted within a scoring model. The final score was calculated as follows: 70% x financial strength score + 30% x impact score. The companies with the highest score were awarded as "World's Top EdTech Companies 2025" by TIME and Statista.Alongside the list of 350 top EdTech companies, Statista and TIME are also publishing a “Rising Stars Ranking” highlighting the companies with the highest revenue growth rates over the past three years out of those who submitted company data via the online application process. Applied criteria:company age 10 years and youngerminimum revenue of 1 million U.S. dollars in 2022revenue growth rate of more than 20% over the last three yearsDisclaimerThe selection of the companies and the definition of the evaluation criteria were based on independent journalistic criteria of TIME and Statista. The evaluation was carried out by the statistics and market research company Statista. TIME and Statista make no claim to the completeness of the companies examined.The ranking only includes companies that qualify according to the criteria described in this document. A position in the ranking is a positive recognition based on research of publicly available data sources at the time and information received via an application process and through collaboration partners.In addition, events after March 14, 2025, were not considered in the analysis. As such, the results of this list should not be used as the sole source of information for future deliberations. The information provided through this ranking should rather be considered in conjunction with other available information about a company.The quality of companies that are not included in the ranking is not disputed.

Donald Trump wants the U.S. to be a leader in artificial intelligence. In January, he signed an executive order intended to enhance America’s "dominance" in AI. In early April, his Administration directed every federal agency to find and hire more people with experience designing and deploying artificial intelligence. On Wednesday, he signed yet another executive order on AI, this one about integrating it into the nation's schools. “AI is where it seems to be at," Trump said.But Trump’s erratic purge of the federal workforce has undermined those very efforts. The Biden Administration moved aggressively in its final 18 months to convince more than 200 AI technology experts to forgo the private sector for the federal workforce, through what was called the ”National AI Talent Surge.” The new hires were deployed throughout the government and used AI to find ways to reduce Social Security wait times, simplify tax filings, and help veterans track their medical care. Most of them were quickly pushed out by the new administration, multiple former federal officials tell TIME. The shift, say the former officials, represents an enormous waste of federal resources, as agencies across the Trump Administration are looking to draw workers with the very experience they just let go. It also means agencies may have to increasingly rely on costlier outside companies for that expertise. The White House and the Office of Management and Budget did not respond to requests for comment.Much of the loss of those AI experts came about when Elon Musk’s Department of Government Efficiency fired hundreds of recent technology hires as part of its broader termination of thousands of employees on probation or so-called “term” hires, former officials who worked in those offices said. Others were fired when Musk’s team subsumed the U.S. Digital Service and when Musk eliminated a technology office at the General Services Administration. That office, called 18F, had helped various government agencies spin up new services, including the IRS’ popular free tax-filing program Direct File. Federal agencies are routinely on the lookout for tech workers, whose skills are in high demand and who can often draw far better salaries in the private sector. Hiring AI experts into government has been a major challenge, says Julie Siegel, who was a senior official in Biden’s Office of Management and Budget. “Everybody is trying to hire AI specialists, so AI was really hard, but we did this big push,” Siegel says.The Trump administration has laid out its own ambitious goals for recruiting more tech talent. On April 3, Russell Vought, Trump’s Director of the Office of Management and Budget, released a 25-page memo for how federal leaders were expected to accelerate the government’s use of AI. “Agencies should focus recruitment efforts on individuals that have demonstrated operational experience in designing, deploying, and scaling AI systems in high-impact environments,” Vought wrote.Putting that into action will be harder than it needed to be, says Deirdre Mulligan, who directed the National Artificial Intelligence Initiative Office in the Biden White House. “The Trump Administration’s actions have not only denuded the government of talent now, but I’m sure that for many folks, they will think twice about whether or not they want to work in government,” Mulligan says. “It’s really important to have stability, to have people’s expertise be treated with the level of respect it ought to be and to have people not be wondering from one day to the next whether they’re going to be employed.”In early 2024, Biden officials hired a new recruiter in Angelica Quirarte, who had spent years pitching tech experts on becoming public servants. She says that coders and engineers are natural problem-solvers and are attracted to the challenge of working with huge data sets that can improve services for millions of people. Previously she had built and led a non-profit national tech talent search called “Tech to Gov” that recruited hundreds of technologists to work in federal and state government.In less than a year, Quirarte tells TIME, she helped hire about 250 AI experts. After Trump’s actions, she estimates about 10% of that cohort are still with the federal government.“It’s going to be really hard” for the Trump administration to hire more tech workers after such haphazard layoffs, Quirarte says. “It’s so chaotic.”Quirarte had initially intended to stay on during the Trump Administration and continue recruiting tech workers into federal service. She had previously spent years in senior roles in California state government under different administrations. “I think transitions are healthy for democracies, when they’re approached with good intent and honor, and most of my work is not political,” Quirarte says.After 23 days in the Trump Administration, Quirarte decided she had had enough and resigned. “It was not an environment where you assumed good intent—you’re operating out of fear,” she says. “That’s not an environment where you can get good policy and good governing work done.”

After years of false starts and regulatory crackdowns, crypto’s future in America may be reignited by something deceptively simple: a digital dollar that doesn’t fluctuate in value.Stablecoins—cryptocurrencies pegged to the U.S. dollar—have quietly become the centerpiece of Washington’s new approach to digital assets. With President Donald Trump now publicly embracing crypto and bipartisan momentum building in Congress, lawmakers have a rare chance to pass the country’s first major crypto legislation. But industry leaders warn: move too slowly, or get too ambitious, and even that narrow victory could slip away.That was the consensus at a TIME100 Talks conversation titled “The Future of Finance: Can Regulation Power Innovation?” on Saturday, April 26, where some of crypto’s most influential players gathered ahead of the White House Correspondents’ Dinner to dig into the current debate over digital asset regulation that could well shape the next chapter of finance and economic progress for decades to come.The panel was moderated by TIME technology correspondent Andrew R. Chow and featured Dante Disparte, the chief strategy officer at Circle—the sponsor of the event—acting Commodity Futures Trading Commission (CFTC) chair Caroline Pham, former Federal Deposit Insurance Corporation (FDIC) chair Jelena McWilliams, and former White House National Security Council special adviser Carole House. The panelists agreed that a focused, narrow bill regulating stablecoins could help cement the U.S. dollar’s dominance in a digital future. They argued that Congress should pass stablecoin legislation now, before broader fights over crypto market structure derail the opportunity.“Having worked in the Senate on many legislative efforts, it's hard to get a bill passed,” McWilliams said with a chuckle. “You get done what you can get done, and you call it a victory.”But despite the broad consensus on the issue, the road to regulation is tricky. Recent reports detail how some crypto companies and lobbyists have pushed to pair stablecoin legislation with a broader, more complex market structure bill—an approach the panelists appeared to oppose. “It would be an utter mistake to try to create a joint and severable model where stablecoins and market structure go side by side,” Disparte said. “It would deny the President a bipartisan legislative win… The stablecoin bill is ready to go and the President wants it on his desk to sign. They can do it before the summer recess, and there's energy—and I think a deep national security imperative to get it right.”Indeed, there appears to be a growing energy around crypto in Washington. Trump, who once called Bitcoin a “scam,” has embraced the industry—and the industry has embraced him right back. During the 2024 election cycle, crypto companies and executives poured tens of millions into campaign donations, helping usher in a Congress far more friendly to digital assets. In March, Trump said that he hoped to sign stablecoin legislation by August. Since then, both the House and Senate have advanced bills—the STABLE Act and the GENIUS Act—out of committee, setting the stage for what could be the first major piece of crypto legislation in U.S. history.At its core, a stablecoin is a digital token pegged to a fiat currency like the U.S. dollar, designed to offer the speed and accessibility of crypto without the volatility of assets like Bitcoin or Ethereum. Today, stablecoins represent a $238 billion market—up from $152 billion just a year ago—and are increasingly used for everything from cross-border payments to decentralized finance. Supporters argue that they could bolster the dollar’s global standing, provide lifelines to inflation-stricken economies, and modernize the U.S. financial system.“To me, the key question here is not really the vehicle, but we have to focus on the outcome and the prize,” Pham said. “The prize here is that good regulations are good. Anybody who thinks that you're going to [have a] money or banking system with no regulations is ridiculous.”Stablecoin regulation, the panelists argued, isn’t just about innovation. It’s also about financial stability, national security, and global competitiveness. “We need to be able to provide for instant payment at the speed of the Internet and at the speed of human need,” Disparte said. “Domestically, it's a misery to move money fast.”House agreed that the moment to act had finally arrived, noting that stablecoin legislation has been debated for three consecutive Congresses, stalling each time over partisan divisions and industry infighting. “This has been the low-hanging fruit,” she said. “I'm delighted that we're finally going to get it, but I agree that it would be a mistake [to pass both bills together].”Still, not everything about the pending bills inspires confidence. Some panelists raised alarms about a provision in the Senate’s GENIUS Act that would allow foreign stablecoin issuers to apply for U.S. licenses directly through the Treasury Department—without the same oversight faced by domestic firms. That, they warned, could open the door to major players like Tether, an international cryptocurrency whose opaque reserve practices and regulatory controversies have long worried U.S. officials.Asked by Chow, the author of Cryptomania, whether a foreign issuer should be able to put out a stablecoin in the U.S. without direct supervision of the U.S. authorities, Disparte warned that letting foreign issuers operate under looser rules could weaken trust in U.S. dollar-backed assets. “I cannot build a car without an airbag and then drop it on the streets of the United States without complying to our safety standards,” he said. “What's the incentive to stay in the U.S. if a bill is passed that allows an offshore issuer to have free access to the US banking system? … Wouldn't it be nice if the rest of this largely amorphous industry also planted flags in the United States? You should not be able to counterfeit the U.S. dollar anywhere in the world and not obligate yourself to U.S. law and U.S. competitiveness.”House also stressed the dangers of creating an uneven playing field for U.S.-based firms, warning that if Congress creates a system where offshore stablecoin issuers have easier access than U.S. firms, that is “not, in fact, setting the stage for true competitiveness for U.S. markets.”Still, political complications loom. Trump’s own growing entanglement with the crypto industry—including the launch of a Trump-branded stablecoin through World Liberty Financial, the crypto venture backed by his family—has already sparked ethical concerns. Critics worry that Trump’s direct financial stake in a market he’s trying to regulate could complicate legislative efforts—or worse, politicize them.Pham, who was nominated by former President Joe Biden as CFTC commissioner and then designated acting chair by Trump, noted the current Administration has made crypto a much larger focus. “The previous Administration had been quite skeptical of the crypto industry, if not downright hostile,” she said. In contrast, she praised the Trump Administration’s openness, citing the first-ever White House Crypto Summit last month—just a day Trump signed an Executive Order establishing a "Strategic Bitcoin Reserve."“It was the first time to be there in the room with these new market participants in an open and welcoming way without stigma,” Pham said. She emphasized the need for “a principles-based regulatory framework that allows for fair competition and responsible innovation,” adding that the CFTC had already rescinded crypto-related advisories from 2018 to reflect how much the market has evolved.McWilliams, the former FDIC chair, echoed that sentiment: “The crypto revolution, or cryptopalooza, is happening,” she said, adding that “if you're not at the table, you're on the table.”TIME100 Talks - The Future of Finance: Can Regulation Power Innovation? was presented by Circle.

As the AI boom continues to escalate, states have rushed to entice data centers with juicy tax breaks. Data centers not only provide cloud based storage, but also power the training of increasingly powerful AI models. Lawmakers in more than 30 states have carved out tax incentives for data center companies, arguing that without them, the data centers wouldn’t come—and that their presence is essential toward growing property and income tax revenue, and driving economic development. But some economists and policy experts have started to question this logic. In a new study from Good Jobs First, a nonprofit research group, authors Greg LeRoy and Kasia Tarczynska find that data center tax breaks have swelled to billions of dollars in lost revenue for states a year—and that those losses, for some states, actually outweigh the tax revenue that the data centers bring in.“It’s a billion-dollar subsidy to the industry that is dominated by a few big tech companies,” says Tarczynska. Conversely, other economists and data center lobbyists argue that the industry’s overall impact vastly outweighs those potential losses, including by supporting local infrastructure and catalyzing further investment. The paper’s authors contend that even if there are positives, the public doesn’t yet know the extent of their costs, which are often obscured by tax privacy laws, confidentiality agreements or a sheer lack of research. “All politicians and corporations want to talk about are the benefits, while ignoring or downplaying the costs of development subsidies,” LeRoy writes in an email to TIME. “Let the companies compete on quality and price, not on how much they can extract from taxpayers and distort economic development and land use.” ‘Up and up and up’Many data center agreements are shrouded in secrecy, which means that the paper’s authors had limited data to draw from. At least 12 states, Leroy says, share no data on tax expenditures for data centers at all. But by looking through state budgets and other fiscal reports, the paper’s authors found that at least 10 states are missing out on $100 million per year in tax revenue. Much of this comes from subsidies for equipment like expensive servers, which need to be swapped out every few years. Many of these exemptions are uncapped in terms of their dollar amounts or time limit. This means that as data centers grow in size and power, so do the subsidies, forcing states to drastically recalculate their budget projections. In Texas, the cost projection for its data center sales tax exemption program increased from $157 million in 2023 to $1 billion this year. “These hyperscalers come and spend a lot of money, and the tax breaks are going up and up and up,” says Tarczynska. Data center proponents offer several rebuttals. First, they argue that most states already offer tax exemptions for manufacturing equipment—and these new carveouts are simply bringing the data center industry in line with those rules. They also argue that without subsidies, data center companies would set up shop elsewhere. In 2022, a tax incentive evaluation study from the Carl Vinson Institute of Government at the University of Georgia found that 90% of data center activity in Georgia was due to the incentive—implying that without those tax breaks, all of that business would essentially vanish. However, there is debate in economic circles around how much incentives actually cause economic activity. A 2018 paper from the economist Timothy Bartik found that the percentage of firms who made a location decision based on an incentive was much lower than the Vinson number: between 2 and 25 percent. “You need to allow for the fact that for some incentives, you would have gotten some data center activity anyway,” Bartik tells TIME. “My view is that there are potentially some fiscal benefits of data centers, but the industry is vastly overstating them.”A debate in GeorgiaData centers are a particular point of contention in Georgia, where they are growing at an unprecedented rate, particularly in the Atlanta metro area. State reports show that data center incentives have waived at least $163 million in local state and sales tax collections each year since 2022. The Governor’s Office of Planning and Budget estimates those incentives will surpass $327 million next year. Last year, Georgia state senate finance committee chairman Chuck Hufstetler, a Republican, spearheaded efforts to scrutinize tax breaks, resulting in the state legislature passing a bill halting tax breaks to data centers for two years. But Governor Brian Kemp vetoed it, saying that it would harm investment plans. Hufstetler tells TIME that the incentives are unnecessary. “The thing I’ve seen is that it's a competition of who can build it the quickest. It’s more important to them that they’re built first versus the cost,” he says. “I’m not opposed to data centers. But if they’re going to come, it shouldn’t be on the backs of local people, and they should be paying all their costs.“ Georgia state senator Sally Harrell, a Democrat, agrees, and contends that data centers do not provide long term jobs or serve local businesses. Data centers typically employ just a few dozen workers, which pales in comparison to factories of similar sizes. “The local community is left footing the bill for big, national companies that generate economic activity elsewhere,” she writes in an email to TIME. The aforementioned 2022 Vinson study in Georgia found that the amount of revenue forgone due to exemptions far exceeded the tax revenue generated by data centers, resulting in a negative state fiscal impact of $18 million in 2021. The study argued, however, that because the data centers would mostly not have been established without the incentive, their overall economic impact was significantly positive.Nathan Jensen, a professor at the University of Texas Austin, argues that this line of reasoning is flawed. “It’s always a bit of a trick where they cite the economic impact: You could go to a Walmart parking lot and throw $100 in the air, and that would have a net positive economic impact,” he says. “The key is the fiscal impact: does the state make its money back?” A defense of data centersStephen Hartka, vice president of research at the Virginia Economic Development Partnership, has a different perspective. He says that the “overwhelming majority” of data centers in Virginia were brought to the state by incentives. He argues that economists who only look at state tax revenues miss the fact that Virginia’s tax structure sends more benefits to smaller localities—who especially benefit from the property taxes that data centers pay. He also says that data centers can be particularly valuable in rural areas that have struggled to attract businesses. “Data centers are a very good fit for these rural areas, because they don't require a lot of people and they generate a ton of revenue for local governments,” he says. “They are financial boons to these communities that might not have many other options.” However, there has been a groundswell of activism against data centers in many Virginian communities. Last year, Ian Lovejoy, a Republican state delegate in Virginia, told TIME that data centers were the “number one issue we hear about from our constituents.” Residents expressed concerns about data centers threatening electricity and water access, and worried that taxpayers may have to foot the bill for future power lines. Across the country, communities have complained about data centers emitting pollutants and harming quality of life. LeRoy, the author of the Good Jobs First study, cautions localities from making “land use decisions based not on jobs or quality of life, but on how much they can get in tax revenue from a particular use.” He points to a precursor: when localities rushed to build malls and big-box retail centers for the sales and property taxes they brought in, only for them to be deserted when e-commerce took its place. The direction of data centers and their national expansion is still hazy. But watchdogs like LeRoy and policy experts like Jensen wish that more data was provided about the massive tradeoffs that states are making, and whether they are worth it. “I think the simple story is these are heavily subsidized capital intensive investments, which makes a pretty tough value proposition for communities,” Jensen says. “Because it's the tax revenues that are the biggest benefit, and you're giving a ton of that back.” 

In the latest development in an ongoing struggle over OpenAI's future direction—and potentially the future of artificial intelligence itself—dozens of prominent figures are urging the Attorneys General of California and Delaware to block OpenAI’s controversial plan to convert from its unique nonprofit-controlled structure to a for-profit company. In a letter made public April 23, signatories including “AI Godfather” Geoffrey Hinton, Harvard legal professor Lawrence Lessig, and several former OpenAI researchers argue the move represents a fundamental betrayal of OpenAI’s founding mission. “The proposed restructuring would eliminate essential safeguards, effectively handing control of, and profits from, what could be the most powerful technology ever created to a for-profit entity with legal duties to prioritize shareholder returns,” the letter’s authors write. It lands as OpenAI faces immense pressure from the other side: failing to implement the restructure by the end of the year could cost the company $20 billion and hamstring future fundraising.OpenAI was founded in 2015 as a non-profit, with its stated mission being to ensure that artificial general intelligence (AGI) “benefits all of humanity" rather than advancing "the private gain of any person." AGI, which OpenAI defines as systems outperforming humans at most economically valuable work, was seen as potentially world-changing but also carrying clear risks, especially if controlled solely by a for-profit company. By 2019, believing they’d need to attract outside investment to build AGI, OpenAI’s leadership created a “capped-profit” subsidiary controlled by the original nonprofit—a hybrid that has allowed the firm to take on over $60 billion in capital over the years to become one of the most valuable startups in history. CEO Sam Altman himself testified to Congress in 2023 that this structure "ensures it remains focused on [its] long-term mission." Then, in December, OpenAI proposed dismantling that unique arrangement, morphing its capped-profit arm into a public benefit corporation, which would take control of OpenAI’s operations and business. The original nonprofit, while relinquishing direct control, would become—through owning a significant equity in the new company—a massively endowed foundation; it would hire its own leadership to fund and pursue separate charitable work in fields such as science and education. OpenAI says the new arrangement would enable them to “raise the necessary capital with conventional terms like others in this space.” Indeed, the need for such terms appears already baked into recent deals: investors from OpenAI’s most recent $40 billion fundraising round, finalized in March, can withdraw half that amount if OpenAI doesn’t restructure by the end of this year.“Our Board has been very clear: our nonprofit will be strengthened and any changes to our existing structure would be in service of ensuring the broader public can benefit from AI. Our for-profit will be a public benefit corporation, similar to several other AI labs like Anthropic - where some of these former employees now work - and xAI, except that they do not support a nonprofit,” an OpenAI spokesperson told TIME via email. “This structure will continue to ensure that as the for-profit succeeds and grows, so too does the nonprofit, enabling us to achieve the mission.”Under the restructure, board members would still legally have to consider OpenAI’s founding mission—albeit it would be downgraded, having to be weighed against profits. “The nonprofit has the authority to basically shut down the company if it thinks it's deviating from [OpenAI’s] mission. Think of it as an off-switch,” Stuart Russell tells TIME. Russell is one of the letter's signatories and a UC Berkeley computer science professor, who co-authored the field's standard textbook. “Basically, they're proposing to disable that off-switch,” he says.That OpenAI’s competitors are for-profit is besides the point, says Sunny Gandhi, vice president of political affairs at youth-led advocacy group Encode Justice and one of the letter’s signatories. “It’s sort of like asking a conservation nonprofit why they can't convert to a logging company just because there are other logging companies out there,” he says. “I think that it would be great if xAI and Anthropic were also nonprofit, but they're not,” he adds. If OpenAI wants to prioritize competitiveness over its original mission, Gandhi says “that's the problem that their original structure was trying to prevent.”The open letter’s targeting of the Attorneys General Rob Bonta of California and Kathy Jennings of Delaware is strategic. In March, Elon Musk lost his bid for an immediate preliminary injunction that would block OpenAI’s conversion, but the decision turned largely on Musk's questionable legal standing—or interest in the case—not the conversion's inherent legality. The judge indicated Musk’s argument that the for-profit shift breaches OpenAI's charitable charter is worthy of further consideration, expediting the trial to this fall. Unlike Musk, however, California and Delare’s Attorneys General have a clear legal interest in the case.California’s Attorney General Rob Bota’s office is reportedly already investigating OpenAI’s plans, and Delaware Attorney General Kathy Jennings has previously signalled she intends to scrutinize any restructuring. Neither responded to TIME’s request for comment on the letter specifically. But how they act may set a precedent, signaling whether corporate governance structures designed to preserve a company’s ideals can withstand the financial gravity of the AI gold rush, or will ultimately buckle under its weight.

A new study claims that AI models like ChatGPT and Claude now outperform PhD-level virologists in problem-solving in wet labs, where scientists analyze chemicals and biological material. This discovery is a double-edged sword, experts say. Ultra-smart AI models could help researchers prevent the spread of infectious diseases. But non-experts could also weaponize the models to create deadly bioweapons. The study, shared exclusively with TIME, was conducted by researchers at the Center for AI Safety, MIT’s Media Lab, the Brazilian university UFABC, and the pandemic prevention nonprofit SecureBio. The authors consulted virologists to create an extremely difficult practical test which measured the ability to troubleshoot complex lab procedures and protocols. While PhD-level virologists scored an average of 22.1% in their declared areas of expertise, OpenAI’s o3 reached 43.8% accuracy. Google's Gemini 2.5 Pro scored 37.6%. Seth Donoughe, a research scientist at SecureBio and a co-author of the paper, says that the results make him a “little nervous,” because for the first time in history, virtually anyone has access to a non-judgmental AI virology expert which might walk them through complex lab processes to create bioweapons. “Throughout history, there are a fair number of cases where someone attempted to make a bioweapon—and one of the major reasons why they didn’t succeed is because they didn’t have access to the right level of expertise,” he says. “So it seems worthwhile to be cautious about how these capabilities are being distributed.”Months ago, the paper’s authors sent the results to the major AI labs. In response, xAI published a risk management framework pledging its intention to implement virology safeguards for future versions of its AI model Grok. OpenAI told TIME that it "deployed new system-level mitigations for biological risks" for its new models released last week. Anthropic included model performance results on the paper in recent system cards, but did not propose specific mitigation measures. Google’s Gemini declined to comment to TIME.AI in biomedicineVirology and biomedicine have long been at the forefront of AI leaders’ motivations for building ever-powerful AI models. “As this technology progresses, we will see diseases get cured at an unprecedented rate,” OpenAI CEO Sam Altman said at the White House in January while announcing the Stargate project. There have been some encouraging signs in this area. Earlier this year, researchers at the University of Florida’s Emerging Pathogens Institute published an algorithm capable of predicting which coronavirus variant might spread the fastest.But up to this point, there had not been a major study dedicated to analyzing AI models’ ability to actually conduct virology lab work. “We've known for some time that AIs are fairly strong at providing academic style information,” says Donoughe. “It's been unclear whether the models are also able to offer detailed practical assistance. This includes interpreting images, information that might not be written down in any academic paper, or material that is socially passed down from more experienced colleagues.”So Donoughe and his colleagues created a test specifically for these difficult, non-Google-able questions. “The questions take the form: ‘I have been culturing this particular virus in this cell type, in these specific conditions, for this amount of time. I have this amount of information about what's gone wrong. Can you tell me what is the most likely problem?’” Donoughe says. And virtually every AI model outperformed PhD-level virologists on the test, even within their own areas of expertise. The researchers also found that the models showed significant improvement over time. Anthropic’s Claude 3.5 Sonnet, for example, jumped from 26.9% to 33.6% accuracy from its June 2024 model to its October 2024 model. And a preview of OpenAI’s GPT 4.5 in February outperformed GPT-4o by almost 10 percentage points. “Previously, we found that the models had a lot of theoretical knowledge, but not practical knowledge,” Dan Hendrycks, the director of the Center for AI Safety, tells TIME. “But now, they are getting a concerning amount of practical knowledge.” Risks and rewardsIf AI models are indeed as capable in wet lab settings as the study finds, then the implications are massive. In terms of benefits, AIs could help experienced virologists in their critical work fighting viruses. Tom Inglesby, the director of the Johns Hopkins Center for Health Security, says that AI could assist with accelerating the timelines of medicine and vaccine development and improving clinical trials and disease detection. “These models could help scientists in different parts of the world, who don't yet have that kind of skill or capability, to do valuable day-to-day work on diseases that are occurring in their countries,” he says. For instance, one group of researchers found that AI helped them better understand hemorrhagic fever viruses in sub-Saharan Africa. But bad-faith actors can now use AI models to walk them through how to create viruses—and will be able to do so without any of the typical training required to access a Biosafety Level 4 (BSL-4) laboratory, which deals with the most dangerous and exotic infectious agents. “It will mean a lot more people in the world with a lot less training will be able to manage and manipulate viruses,” Inglesby says. Hendrycks urges AI companies to put up guardrails to prevent this type of usage. “If companies don't have good safeguards for these within six months time, that, in my opinion, would be reckless,” he says. Hendrycks says that one solution is not to shut these models down or slow their progress, but to make them gated, so that only trusted third parties get access to their unfiltered versions. “We want to give the people who have a legitimate use for asking how to manipulate deadly viruses—like a researcher at the MIT biology department—the ability to do so,” he says. “But random people who made an account a second ago don't get those capabilities.” And AI labs should be able to implement these types of safeguards relatively easily, Hendrycks says. “It’s certainly technologically feasible for industry self-regulation,” he says. “There’s a question of whether some will drag their feet or just not do it.” xAI, Elon Musk’s AI lab, published a risk management framework memo in February, which acknowledged the paper and signaled that the company would “potentially utilize” certain safeguards around answering virology questions, including training Grok to decline harmful requests and applying input and output filters. OpenAI, in an email to TIME on Monday, wrote that its newest models, the o3 and o4-mini, were deployed with an array of biological-risk related safeguards, including blocking harmful outputs. The company wrote that it ran a thousand-hour red-teaming campaign in which 98.7% of unsafe bio-related conversations were successfully flagged and blocked. "We value industry collaboration on advancing safeguards for frontier models, including in sensitive domains like virology," a spokesperson wrote. "We continue to invest in these safeguards as capabilities grow."Inglesby argues that industry self-regulation is not enough, and calls for lawmakers and political leaders to strategize a policy approach to regulating AI’s bio risks. “The current situation is that the companies that are most virtuous are taking time and money to do this work, which is good for all of us, but other companies don't have to do it,” he says. “That doesn't make sense. It's not good for the public to have no insights into what's happening.”“When a new version of an LLM is about to be released,” Inglesby adds, “there should be a requirement for that model to be evaluated to make sure it will not produce pandemic-level outcomes.”

Tech companies are investing hundreds of billions of dollars to build new U.S. datacenters where —if all goes to plan—radically powerful new AI models will be brought into existence.But all of these datacenters are vulnerable to Chinese espionage, according to a report published Tuesday. At risk, the authors argue, is not just tech companies’ money, but also U.S. national security amid the intensifying geopolitical race with China to develop advanced AI.The unredacted report was circulated inside the Trump White House in recent weeks, according to its authors. TIME viewed a redacted version ahead of its public release. The White House did not respond to a request for comment.Today’s top AI datacenters are vulnerable to both asymmetrical sabotage—where relatively cheap attacks could disable them for months—and exfiltration attacks, in which closely guarded AI models could be stolen or surveilled, the report’s authors warn.Even the most advanced datacenters currently under construction—including OpenAI’s Stargate project—are likely vulnerable to the same attacks, the authors tell TIME.“You could end up with dozens of datacenter sites that are essentially stranded assets that can’t be retrofitted for the level of security that’s required,” says Edouard Harris, one of the authors of the report. “That’s just a brutal gut-punch.”The report was authored by brothers Edouard and Jeremie Harris of Gladstone AI, a firm that consults for the U.S. government on AI’s security implications. In their year-long research period, they visited a datacenter operated by a top U.S. technology company alongside a team of former U.S. special forces who specialize in cyberespionage.In speaking with national security officials and datacenter operators, the authors say, they learned of one instance where a top U.S. tech company’s AI datacenter was attacked and intellectual property was stolen. They also learned of another instance where a similar datacenter was targeted in an attack against a specific unnamed component which, if it had been successful, would have knocked the entire facility offline for months.The report addresses calls from some in Silicon Valley and Washington to begin a “Manhattan Project” for AI, aimed at developing what insiders call superintelligence: an AI technology so powerful that it could be used to gain a decisive strategic advantage over China. All the top AI companies are attempting to develop superintelligence—and in recent years both the U.S. and China have woken up to its potential geopolitical significance.Although hawkish in tone, the report does not advocate for or against such a project. Instead, it says that if one were to begin today, existing datacenter vulnerabilities could doom it from the start. “There's no guarantee we'll reach superintelligence soon,” the report says. “But if we do, and we want to prevent the [Chinese Communist Party] from stealing or crippling it, we need to start building the secure facilities for it yesterday.”China Controls Key Datacenter PartsMany critical components for modern datacenters are mostly or exclusively built in China, the report points out. And due to the booming datacenter industry, many of these parts are on multi-year back orders.What that means is that an attack on the right critical component can knock a datacenter offline for months—or longer.Some of these attacks, the report claims, can be incredibly asymmetric. One such potential attack—the details of which are redacted in the report—could be carried out for as little as $20,000, and if successful could knock a $2 billion datacenter offline from between six months to a year.China, the report points out, is likely to delay shipment of components necessary to fix datacenters brought offline by these attacks, especially if it considers the U.S. to be on the brink of developing superintelligence. “We should expect that the lead times on China-sourced generators, transformers, and other critical data center components will start to lengthen mysteriously beyond what they already are today,” the report says. “This will be a sign that China is quietly diverting components to its own facilities, since after all, they control the industrial base that is making most of them.”AI Labs Struggle With Basic Security, Insiders WarnThe report says that neither existing datacenters nor AI labs themselves are secure enough to prevent AI model weights—essentially their underlying neural networks—from being stolen by nation-state level attackers.The authors cite a conversation with a former OpenAI researcher who described two vulnerabilities that would allow attacks like that to happen—one of which had been reported on the company’s internal Slack channels, but was left unaddressed for months. The specific details of the attacks are not included in the version of the report viewed by TIME.An OpenAI spokesperson said in a statement: “It’s not entirely clear what these claims refer to, but they appear outdated and don’t reflect the current state of our security practices. We have a rigorous security program overseen by our Board’s Safety and Security Committee.”The report's authors acknowledge that things are slowly getting better. “According to several researchers we spoke to, security at frontier AI labs has improved somewhat in the past year, but it remains completely inadequate to withstand nation state attacks,” the report says. “According to former insiders, poor controls at many frontier AI labs originally stem from a cultural bias towards speed over security.”Independent experts agree many problems remain. "There have been publicly disclosed incidents of cyber gangs hacking their way to the [intellectual property] assets of Nvidia not that long ago," Greg Allen, the director of the Wadhwani AI Center at the Washington think-tank the Center for Strategic and International Studies, tells TIME in a message. "The intelligence services of China are far more capable and sophisticated than those gangs. There’s a bad offense / defense mismatch when it comes to Chinese attackers and U.S. AI firm defenders."Superintelligent AI May Break FreeA third crucial vulnerability identified in the report is the susceptibility of datacenters—and AI developers—to powerful AI models themselves.In recent months, studies by leading AI researchers have shown top AI models beginning to exhibit both the drive, and the technical skill, to “escape” the confines placed on them by their developers.In one example cited in the report, during testing, an OpenAI model was given the task of retrieving a string of text from a piece of software. But due to a bug in the test, the software didn’t start. The model, unprompted, scanned the network in an attempt to understand why—and discovered a vulnerability on the machine it was running on. It used that vulnerability, also unprompted, to break out of its test environment and recover the string of text that it had initially been instructed to find.“As AI developers have built more capable AI models on the path to superintelligence, those models have become harder to correct and control,” the report says. “This happens because highly capable and context-aware AI systems can invent dangerously creative strategies to achieve their internal goals that their developers never anticipated or intended them to pursue.”The report recommends that any effort to develop superintelligence must develop methods for “AI containment,” and allow leaders with a responsibility for developing such precautions to block the development of more powerful AI systems if they judge the risk to be too high.“Of course,” the authors note, “if we’ve actually trained a real superintelligence that has goals different from our own, it probably won’t be containable in the long run.”

This story is part of the 2025 TIME100. Read Jennifer Doudna’s tribute to Demis Hassabis here. Demis Hassabis learned he had won the 2024 Nobel Prize in Chemistry just 20 minutes before the world did. The CEO of Google DeepMind, the tech giant’s artificial intelligence lab, received a phone call with the good news at the last minute, after a failed attempt by the Nobel Foundation to find his contact information in advance. “I would have got a heart attack,” Hassabis quips, had he learned about the prize from the television. Receiving the honor was a “lifelong dream,” he says, one that “still hasn’t sunk in” when we meet five months later.Hassabis received half of the award alongside a colleague, John Jumper, for the design of AlphaFold: an AI tool that can predict the 3D structure of proteins using only their amino acid sequences—something Hassabis describes as a “50-year grand challenge” in the field of biology. Released freely by Google DeepMind for the world to use five years ago, AlphaFold has revolutionized the work of scientists toiling on research as varied as malaria vaccines, human longevity, and cures for cancer, allowing them to model protein structures in hours rather than years. The Nobel Prizes in 2024 were the first in history to recognize the contributions of AI to the field of science. If Hassabis gets his way, they won’t be the last.AlphaFold’s impact may have been broad enough to win its creators a Nobel Prize, but in the world of AI, it is seen as almost hopelessly narrow. It can model the structures of proteins but not much else; it has no understanding of the wider world, cannot carry out research, nor can it make its own scientific breakthroughs. Hassabis’s dream, and the wider industry’s, is to build AI that can do all of those things and more, unlocking a future of almost unimaginable wonder. All human diseases will be a thing of the past if this technology is created, he says. Energy will be zero-carbon and free, allowing us to transcend the climate crisis and begin restoring our planet’s ecosystems. Global conflicts over scarce resources will dissipate, giving way to a new era of peace and abundance. “I think some of the biggest problems that face us today as a society, whether that's climate or disease, will be helped by AI solutions,” Hassabis says. “I'd be very worried about society today if I didn't know that something as transformative as AI was coming down the line.”This hypothetical technology—known in the industry as Artificial General Intelligence, or AGI—had long been seen as decades away. But the fast pace of breakthroughs in computer science over the last few years has led top AI scientists to radically revise their expectations of when it will arrive. Hassabis predicts AGI is somewhere between five and 10 years away—a rather pessimistic view when judged by industry standards. OpenAI CEO Sam Altman has predicted AGI will arrive within Trump’s second term, while Anthropic CEO Dario Amodei says it could come as early as 2026. Partially underlying these different predictions is a disagreement over what AGI means. OpenAI’s definition, for instance, is rooted in cold business logic: a technology that can perform most economically valuable tasks better than humans can. Hassabis has a different bar, one focused instead on scientific discovery. He believes AGI would be a technology that could not only solve existing problems, but also come up with entirely new explanations for the universe. A test for its existence might be whether a system could come up with general relativity with only the information Einstein had access to; or if it could not only solve Photograph by David Vintiner for TIMEOrder your copy of the 2025 TIME100 issue hereIn an AI industry whose top ranks are populated mostly by businessmen and technologists, that identity sets Hassabis apart. Yet he must still operate in a system where market logic is the driving force. Creating AGI will require hundreds of billions of dollars’ worth of investments—dollars that Google is happily plowing into Hassabis’ DeepMind unit, buoyed by the promise of a technology that can do anything and everything. Whether Google will ensure that AGI, if it comes, benefits the world remains to be seen; Hassabis points to the decision to release AlphaFold for free as a symbol of its benevolent posture. But Google is also a company that must legally act in the best interests of its shareholders, and consistently releasing expensive tools for free is not a long-term profitable strategy. The financial promise of AI—for Google and for its competitors—lies in controlling a technology capable of automating much of the labor that drives the more than $100 trillion global economy. Capture even a small fraction of that value, and your company will become one of the most profitable the world has ever seen. Good news for shareholders, but bad news for regular workers who may find themselves suddenly unemployed.So far, Hassabis has successfully steered Google's multibillion-dollar AI ambitions toward the type of future he wants to see: one focused on scientific discoveries that, he hopes, will lead to radical social uplift. But will this former child chess prodigy be able to maintain his scientific idealism as AI reaches its high-stakes endgame? His track record reveals one reason to be skeptical. When DeepMind was acquired by Google in 2014, Hassabis insisted on a contractual firewall: a clause explicitly prohibiting his technology from being used for military applications. It was a red line that reflected his vision of AI as humanity's scientific savior, not a weapon of war. But multiple corporate restructures later, that protection has quietly disappeared. Today, the same AI systems developed under Hassabis's watch are being sold, via Google, to militaries such as Israel's—whose campaign in Gaza has killed tens of thousands of civilians. When pressed, Hassabis denies that this was a compromise made in order to maintain his access to Google's computing power and thus realize his dream of developing AGI. Instead, he frames it as a pragmatic response to geopolitical reality, saying DeepMind changed its stance after acknowledging that the world had become “a much more dangerous place” in the last decade. “I think we can't take for granted anymore that democratic values are going to win out,” he says. Whether or not this justification is honest, it raises an uncomfortable question: If Hassabis couldn't maintain his ethical red line when AGI was just a distant promise, what compromises might he make when it comes within touching distance?To get to Hassabis’s dream of a utopian future, the AI industry must first navigate its way through a dark forest full of monsters. Artificial intelligence is a dual-use technology like nuclear energy: it can be used for good, but it could also be terribly destructive. Hassabis spends much of his time worrying about risks, which generally fall into two different buckets. One is the possibility of systems that can meaningfully enhance the capabilities of bad actors to wreak havoc in the world; for example, by endowing rogue nations or terrorists with the tools they need to synthesize a deadly virus. Preventing risks like that, Hassabis believes, means carefully testing AI models for dangerous capabilities, and only gradually releasing them to more users with effective guardrails. It means keeping the “weights” of the most powerful models (essentially their underlying neural networks) out of the public’s hands altogether, so that models can be withdrawn from public use if dangers are discovered after release. That’s a safety strategy that Google follows but which some of its competitors, such as DeepSeek and Meta, do not. The second category of risks may seem like science fiction, but they are taken seriously inside the AI industry as model capabilities advance. These are the risks of AI systems acting autonomously— such as a chatbot deceiving its human creators, or a robot attacking the person it was designed to help. Language models like DeepMind’s Gemini are essentially grown from the ground up, rather than written by hand like old-school computer programs, and so computer scientists and users are constantly finding ways to elicit new behaviors from what are best understood as incredibly mysterious and complex artifacts. The question of how to ensure that they always behave and act in ways that are “aligned” to human values is an unsolved scientific problem. Early signs of misaligned behaviors, like strategic lying, have already been identified by researchers working with today’s language models. Those problems are only likely to become more acute as models get better. “How do we ensure that we can stay in charge of those systems, control them, interpret what they're doing, understand them, and put the right guardrails in place that are not movable by very highly capable self-improving systems?” Hassabis says. “That is an extremely difficult challenge.”It’s a devilish technical problem—but what really keeps Hassabis up at night are the political coordination challenges that accompany it. Even if well-meaning companies can make safe AIs, that doesn’t by itself stop the creation and proliferation of unsafe AIs. Stopping that will require international collaboration—something that’s becoming increasingly difficult as western alliances fray and geopolitical tensions between the U.S. and China rise. Hassabis has played a significant role in the three AI summits held by global governments since 2023, and says he would like to see more of that kind of cooperation. He says the U.S. government’s export controls on AI chips, intended to prevent China’s AI industry from surpassing Silicon Valley, are “fine”—but he would prefer to avoid political choices that “end up in an antagonistic kind of situation.” He might be out of luck. As both the U.S. and China have woken up in recent years to the potential power of AGI, the climate of global cooperation —which reached a high watermark with the first AI Safety Summit in 2023—has given way to a new kind of realpolitik. In this new era, with nations racing to militarize AI systems and build up stockpiles of chips, and with a new cold war brewing between the U.S. and China, Hassabis still holds out hope that competing nations and companies can find ways to set aside their differences and cooperate, at least on AI safety. “It’s in everyone’s self-interest to make sure that goes well,” he says. Even if the world can find a way to safely navigate through the geopolitical turmoil of AGI’s arrival, the question of labor automation will rear its head. When governments and companies no longer rely on humans to generate their wealth, what leverage will citizens have left to demand the ingredients of democracy and a comfortable life? AGI might create abundance, but it won’t dispel the incentives for companies and states to amass resources and compete with rivals. Hassabis admits he is better at forecasting technological futures than social and economic ones; he says he wishes more economists would take the possibility of near-term AGI seriously. Still, he thinks it’s inevitable we’ll need a “new political philosophy” to organize society in this world. Democracy, he says, “is not a panacea, by any means,” and might have to give way to “something better.” Hassabis, left, captaining the England under-11s chess team at the age of 9. Courtesy Demis HassabisAutomation, meanwhile, is already on the horizon. In March, DeepMind announced Gemini 2.5, the latest version of its flagship AI model, which outperforms rival models made by OpenAI and Anthropic on many popular metrics. Hassabis is currently hard at work on Project Astra, a DeepMind effort to build a universal digital assistant powered by Gemini. That work, he says, is not intended to hasten labor disruptions, but instead is about building the necessary scaffolding for the type of AI that he hopes will one day make its own scientific discoveries. Still, as research into these AI “agents” progresses, Hassabis says, expect them to be able to carry out increasingly more complex tasks independently. (An AI agent that can meaningfully automate the job of further AI research, he predicts, is “a few years away.”) For the first time, Google is also now using these digital brains to control robot bodies: in March the company announced a Gemini-powered android robot that can carry out embodied tasks like playing tic-tac-toe, or making its human a packed lunch. The tone of the video announcing Gemini Robotics was friendly, but its connotations were not lost on some YouTube commenters: “Nothing to worry [about,] humanity, we are only developing robots to do tasks a 5 year old can do,” one wrote. “We are not working on replacing humans or creating robot armies.”  Hassabis acknowledges the social impacts of AI are likely to be significant. People must learn how to use new AI models, he says, in order to excel professionally in the future and not risk getting left behind. But he is also confident that if we eventually build AGI capable of doing productive labor and scientific research, the world that it ushers into existence will be abundant enough to ensure a substantial increase in quality of life for everybody. “In the limited-resource world which we're in, things ultimately become zero-sum,” Hassabis says. “What I'm thinking about is a world where it's not a zero-sum game anymore, at least from a resource perspective.” Five months after his Nobel Prize, Hassabis’s journey from chess prodigy to Nobel laureate now leads toward an uncertain future. The stakes are no longer just scientific recognition—but potentially the fate of human civilization. As DeepMind's machines grow more capable, as corporate and geopolitical competition over AI intensifies, and as the economic impacts loom larger, Hassabis insists that we might be on the cusp of an abundant economy that benefits everyone. But in a world where AGI could bring unprecedented power to those who control it, the forces of business, geopolitics, and technological power are all bearing down with increasing pressure. If Hassabis is right, the turbulent decades of the early 21st century could give way to a shining utopia. If he has miscalculated, the future could be darker than anyone dares imagine. One thing is for sure: in his pursuit of AGI, Hassabis is playing the highest-stakes game of his life.

31-year-old Vitalik Buterin is one of crypto’s most important figures. As the founder of Ethereum, he pioneered the idea that crypto and blockchains could serve larger purposes beyond money. But Buterin also sought to diminish his own role within the ecosystem, and encouraged his community to think much bigger than their own short-term gains. A new documentary, Vitalik: An Ethereum Story, tells his story, following him around the world as he confronts difficult technical problems and evangelizes a strange new technology that attempts to reorient the world around its decentralized value system. ”In our minds, tech is never neutral: it’s a reflection of the values and blind spots of its creators,” says co-director Chris Temple. “As we head into conversations around AI and crypto, Vitalik models a new kind of leadership compared to the tech leaders that we're used to—the Elon Musks, Mark Zuckerbergs and Jeff Bezos’s of the world—that are that are structured within these centralized organizations and decisionmaking apparatuses.” Temple and co-director Zach Ingrasci filmed Buterin across two years, and much of their footage didn’t make it into the 86-minute film. That includes one scene in Kyiv, Ukraine, in which Buterin plays chess with Mykhailo Fedorov, Ukraine’s Minister of Digital Transformation. The scene, published exclusively by TIME, shows a fascinating peek into Buterin’s ideological leanings, and his acute desire for crypto to have real-world use cases beyond speculation. Buterin traveled to Ukraine in September 2022, six months after Russia’s invasion. While Buterin was born in Russia, he staunchly opposed the invasion and personally donated millions to Ukrainian relief efforts. Thanks in part to his vocal support, almost $100 million in crypto poured into Ukraine in the first couple weeks of the invasion, offering fast relief and easy, direct transactions.The deleted scene shows Fedorov and Buterin talking over a game of chess. In the early days of the invasion, Fedorov tells him, the country’s national bank had banned international transactions, so the Ukrainian government instead used crypto to receive funding and buy weapons and military supplies. “All of the first drones, lots of ammunition, arrived thanks to crypto,” Fedorov tells him. “We saved the lives of hundreds—maybe thousands of our military. So it was highly important.”Buterin responds: “We love Ukraine… For the blockchain community itself, this was the first opportunity to make a real difference with blockchain and cryptocurrency.” Later, when Fedorov points out that Ukrainians continue to live in the war-torn country, Buterin adds: “I think continuing real life means, even if I’m a good person, to show Putin a middle finger.” Ingrasci says the clip didn’t make it into the final cut because the moment it depicts is referenced in other ways. “But I think it's the most important moment for Vitalik in our journey with him, because it’s this real world use case of how crypto can really make the world a better place,” he says. Vitalik: An Ethereum Story is available on VOD on April 15. TIME Studios served as one of the film’s production companies. Andrew R. Chow’s book about crypto and Sam Bankman-Fried, Cryptomania, was published in August.

Announcing his tariffs in the White House Rose Garden last week, President Trump said the move would help reopen shuttered car factories in Michigan and bring various other jobs back to the U.S.“The president wants to increase manufacturing jobs here in the United States of America,” Press Secretary Karoline Leavitt added on Tuesday. “He wants them to come back home.”But rather than enticing companies to create new jobs in the U.S., economists say, the new tariffs—bolstered by recent advancements in artificial intelligence and robotics—could instead increase incentives for companies to automate human labor entirely.“There’s no reason whatsoever to believe that this is going to bring back a lot of jobs,” says Carl Benedikt Frey, an economist and professor of AI & work at Oxford University. “Costs are higher in the United States. That means there’s an even stronger economic incentive to find ways of automating even more tasks.”In other words: when labor costs are low—like they are in Vietnam—it’s usually not worth it for companies to invest in the expensive up-front costs of automating human labor. But if companies are forced to move their labor to more expensive countries, like the U.S., that cost-benefit calculation changes drastically.To be sure, experts note that tariffs may not immediately lead to more automation. Automating manufacturing jobs often requires companies to make significant investments in physical machinery, which tariffs are likely to make more expensive. In a time of economic turmoil, companies also usually hold off on making big capital expenditures. Thus, in the short run, Nobel Prize-winning economist Daron Acemoglu predicts, there is likely to be so much disruption that few companies will invest in automation or much else. But if tariffs persist in the medium term, Acemoglu tells TIME, he expects companies “will have no choice but to bring some of their supply chains back home—but they will do it via AI and robots.”The evidence from the last time Trump imposed tariffs on trading partners, in 2018, shows no major increase in automation as a result. (Those tariffs did in fact lead to job losses in affected industries anyway, a Federal Reserve study found, due to higher production costs and reduced export competitiveness.) But some economists think the 2025 tariffs could be different—incentivizing more automation—because AI and robotics have come a long way since 2018. “Our technological capabilities have improved since the last round of tariffs, particularly because of improvements in AI,” says Frey, the Oxford economist.The rise of roboticsFor years, a major limitation of robots was that they couldn’t adapt to even minor changes in their environments. An industrial robot might be able to carry out a repeatable task in a controlled environment easily—like cutting a car door from a sheet of metal—but for more deft tasks in more complex environments, humans still prevailed.That might not be the case for much longer. Robot “brains” are getting more adaptable, thanks to progress in general AI systems like large language models. Robot bodies are becoming more deft, thanks to investment and research by companies like Boston Dynamics. And robots are getting cheaper to produce over time (although tariffs might temporarily reverse that trend). “It has taken some time, but people have been doing research on taking language models’ ability for common-sense understanding, and applying it to robotics,” says Lucas Hansen, co-founder of CivAI, a non-profit. “It doesn’t require much special effort to apply robots to new purposes now, especially once this technology matures a bit more. So if you’re a mid-sized manufacturing operation, previously you would have had to invest tons of money in R&D to automate everything. But now, it will require a lot less marginal effort.”Acemoglu is more skeptical. Robots, he says, still struggle in complex environments, even if flashy corporate demo videos suggest otherwise. “I wouldn’t be optimistic that it’s a quick problem to be solved,” he says, predicting that flexible robots are at least 10 years away.If tariffs lead to more automation, it’s still unlikely that productivity gains will offset the huge losses stemming from supply chain disruption and added import costs. “The main first-order effect of tariffs is they will make everything less efficient,” says Erik Brynjolfsson, the director of the Digital Economy Lab at Stanford University. “When you throw sand in the gears of supply chains and global trade, we’re all just going to be a little bit poorer.”The Trump Administration has said it wants AI to benefit American workers, rather than replace them. “We refuse to view AI as a purely disruptive technology that will inevitably automate away our labor force,” Vice President JD Vance said in February. “We believe and we will fight for policies that ensure that AI is going to make our workers more productive, and we expect that they will reap the rewards.”But past experiences with new technologies in the workplace suggest that rosy vision is unlikely to come to pass, says Brian Merchant, a labor historian and author of Blood in the Machine. “Historically when there is a downturn, if there is an opportunity to automate, then companies will take it. That doesn’t necessarily mean that you’ll use fewer humans, but it does mean that employers have a chance to break through labor protections and gain more leverage.”

In response to Donald Trump’s escalating tariffs, China retaliated in part by placing export restrictions on a slew of rare earth elements. These powerful materials are crucial to the U.S., because they underpin the creation of weapons, computer chips, and electric cars. China  produces a majority of these rare earth materials—and experts say that the U.S. is years away from building its own supply chain. As the U.S.–China trade war ramps up, rare earths are among the most important pieces of leverage that China controls. There are many reasons why China would not want to shut off U.S. access to rare earths completely, most notably that the country makes a lot of money from exporting them. But if China decides to further choke off its supply, the ripple effects could be extremely painful across many industries, says Lyle Trytten, a critical minerals expert. “The U.S. does not have the means to create the materials it needs to create the devices it survives on,” he says. Rare earth’s importanceThe importance of rare earths has only increased over the years, due to the world’s reliance on ever-powerful computers and its search for cleaner energy. Dysprosium and terbium, for example, are found in smartphones’ vibration units. Neodymium powers the motors of electric vehicles. Tungsten, an ultra-hard metal, is used in ammunition, semiconductor chips, and alloys found in jet engines and deep-drilling rigs. Almost all of these materials are mined and processed by China, which has spent decades aggressively building the infrastructure to do so. As a result, many companies, including Tesla and Apple, source their rare earths from China. Recently, China has not hesitated to wield this dominance as a geopolitical bargaining tool. In 2010, China halted rare-earth exports to Japan amidst rising tensions. Over the past two years, Beijing has imposed curbs on other critical minerals, such as gallium, germanium, and graphite. “It’s pretty predictable now that once the U.S. pulls something—whether it's an export control on a particular technology or a tariff—this is China’s chosen weapon,” says Fabian Villalobos, an engineer at RAND. “Critically, the separation of heavy rare earths from the light rare earths is where China has a dominance, and therefore there’s a vulnerability in the supply chain.”The White House signaled its understanding of the fragility of the current ecosystem when it exempted critical minerals from its tariffs regime this month. But that did not stop China from issuing export controls on seven kinds of rare earth elements, to all countries, on Friday. The decision is not a ban, but it does give Beijing oversight and control over access to the rare earth elements. China said that its export controls will not affect the rare earth supply chain. Crucially, China omitted several of the most-coveted rare earth elements, including neodymium and praseodymium. But the controls show that China is willing to use these materials as a bargaining chip and could escalate their restrictions if tensions increase. “Consider this an opening shot across the bow,” says Trytten. The listed elements also include those found in microchips used for AI—a further indication of the ongoing AI arms race between the two countries. Villalobos says that in the short term, there will likely be a slowdown of rare earth exports as companies apply for licenses to adhere to the export controls. “You might see a temporary dip in exports, and then a ramping up as more companies get their licenses,” he says. But Villalobos says the greater threat to U.S. companies could come afterward, once China starts collecting detailed information about the rare earth market—which then gives China the ability to impose damaging sanctions upon specific companies. That could include U.S. defense companies like Lockheed Martin, which needs rare earths for components in missile systems and fighter jets. “This is the danger: The more information you can gather from exporters, the more you can target specific companies that you don't want getting access to rare earth,” he says. U.S. capacityMany experts have long called for the U.S. to wean itself off of this dependence. Some believe that the solution is to mine rare earths on the moon. Other entrepreneurs have started projects building mines and processing facilities across America. Trump’s tariffs, then, could incentivize these kinds of shifts; to force American companies to build up supply chain resilience. “Maybe it will move the ball on investments, which is one of the big barriers to diversifying critical mineral supply chains,” Villalobos says. But rare earths and other minerals are extremely intensive to process—and the U.S. does not have the infrastructure to scale these efforts quickly, Trytten says. The number of graduates of U.S. mining engineering programs has steadily declined over the last few decades, potentially leading to a lack of expertise. Trytten says that there is danger in rushing new mining projects into production. “The history of our industry in the metal space is that when we try to do things fast, we tend to do them poorly,” he says. Because of these factors, Trytten contends that even if a new wave of mining projects is kickstarted now, they will not come to fruition until long after Trump has left the White House. “Call it eight to 10 years before you have significant new capacity for a lot of these raw materials,” Trytten says. “Can he weather the storm that long?”Other experts say that various other parts of Trump’s tariffs make it hard for them to scale up their state-side infrastructure. On the Rare Earth Exchanges podcast, the entrepreneur Daniel O’Connor said that tariffed materials like steel and aluminum are crucial toward mining and processing. “Let’s not do tariffs on things we need to build our infrastructure,” he said. Rare earths in Greenland?Some have speculated that rare earths play a major role in Donald Trump’s interest in Greenland. Tech giants like Bill Gates and Jeff Bezos have invested in companies prospecting for rare earths there. But extracting resources out of Greenland poses many challenges. “Greenland has very little domestic energy production, and you can find those resources pretty much anywhere,” Trytten says. “There are much easier mining locations than the Arctic.” Regardless of whether Greenland is a viable option, many U.S. companies are now being forced to pursue non-Chinese rare earth options, even if it takes them years to develop. “Think about every automated thing: If you push a button and it moves, it’s probably reliant on some sort of rare earth magnet,” Villalobos says. “Whoever makes that, if they're in the U.S., Japan, or anywhere outside of China, they’re going to feel the impact from this—and they might be potential targets for sanctions in the future.”

Stocks in AI companies were among the biggest losers after President Trump announced sweeping tariffs on foreign trading partners last week, in a sign that those tariffs could be bad news for the industry.The companies at the forefront of the AI industry are currently spending hundreds of billions of dollars on building new datacenters to train AI models. Tariffs will increase the already gargantuan costs of those efforts, analysts say.“The tariffs will make building AI datacenters much more expensive, both because AI servers are largely imported and will face tariffs, at least until supply chains can be rejigged, and because much of the other equipment in datacenters, like the cooling and power infrastructure, is imported as well,” says Chris Miller, author of Chip War.Chips themselves, the key computing hardware inside AI datacenters, are exempt from Trump’s tariffs—but only if they are imported to the U.S. as standalone products. However, most chips are not imported into the U.S. as raw materials; instead, they arrive already-packaged inside products like servers, which are subject to tariffs. Worried AI investors received good news on Monday in a note circulated by analyst Stacy Rasgon, who pointed out that most Nvidia servers are likely to escape the bite of Trump’s tariffs. That’s because most appear to be assembled in Mexico, and therefore benefit from a tariff exemption under a free trade agreement. That’s a “silver lining” to the news, says Rasgon, a semiconductor industry analyst at Bernstein Research. (Nvidia declined to comment.)“I think there are some workarounds to avoid massive tariffs on AI infrastructure in the U.S., which is good because otherwise what’s the whole point of this?” says Rasgon. “We’d be making the U.S. the most expensive place in the world to build out AI infrastructure—that doesn’t sound like a great thing.”But construction materials, computer parts, cooling infrastructure, and power supplies are just some of the costs that are likely to increase as a result of the tariffs. The costs could be so great that companies might consider building datacenters abroad instead of in the U.S., says Lucas Hansen of the Civic AI Security Program, a nonprofit. Datacenters already tend to congregate where power is cheap, he says. “It’s possible that tariffs are one more additional incentive for building those datacenters abroad.”The increased costs of datacenter construction create a “real risk,” Miller says, that the U.S. might begin losing ground to China in the AI race—victory in which is a key foreign policy goal of the Trump Administration. “It has already been a major challenge to build all the datacenter capacity we need” in the U.S. to stay ahead of China, Miller says. “Now datacenter construction will get meaningfully more expensive.”“The short term impact will be significant, and the long term impact is unclear—and companies can't plan for the long term because tariff rates will likely keep changing,” Miller adds.Even if Trump creates more carveouts to ease the pain on the datacenter industry, the changes to the macroeconomic climate wrought by Trump’s trade war might create new headwinds for AI companies. “My bigger worry is more macro now: we go into recession, ad spending falls off, and the hyperscalers in general have less money,” says Rasgon, referring to the tech companies spending heavily on AI. Collapse in demand for AI and datacenters, plus supply chain chaos, might follow, he adds. “This doesn’t feel like a strategy,” Rasgon says of the tariffs. “This is just a grenade.”The increases in datacenter costs will probably make it more expensive for companies to train AI systems in future. But this is unlikely to mean that using AI gets moreresearch by Epoch AI, as a result of algorithmic efficiencies, hardware improvements, and pricing competition. In other words, a year from now, using a given model will probably require significantly less computing power (and therefore money) than it does today. So even if Trump’s tariffs do add to the cost of datacenter components, researchers say AI usage is likely to continue to get cheaper over time.

During the 2024 elections, the crypto industry spent an unprecedented amount of money on campaign donations in the hopes of encouraging Congress to pass pro-crypto legislation. Those efforts seem to have largely succeeded, with pro-crypto legislators filling the halls of Congress. The first crypto-related area theyve decided to tackle? Stablecoin regulation.Stablecoins are cryptocurrencies designed to hold the value of a U.S. dollar. Proponents argue that stablecoins help the U.S. preserve the global importance of the dollar, while allowing people worldwide to transact more freely, cheaply, and securely. Stablecoin usage is growing enormously: its total market cap is around $235 billion, up from $152 billion just a year ago. In March, President Trump said that he hoped to sign stablecoin legislation by August. Congress has responded accordingly: In the past month, both the House and Senate have advanced stablecoin bills out of committee.Heres a brief overview of stablecoins, the proposed legislation, and potential risks.What are stablecoins and who supports them?Stablecoins are somewhat like bank deposits. Typically, a consumer who wants a stablecoin gives a dollar to an issuing company, who mints a stablecoin on a blockchain. The user can then send that stablecoin around the world as a form of payment to anyone who accepts it. Crypto traders like stablecoins because they do not fluctuate in price nearly as much as assets like Bitcoin or Ethereum, making trading more predictable. And many non-traders across the world appreciate stablecoins because they hold their value better than currencies in countries with high inflation, like Argentina and Turkey.In the U.S., stablecoin supporters come from across the political spectrum. On the right, political leaders like House Majority Whip Tom Emmer argue that stablecoins help maintain the dollars status as the worlds reserve currency. A massive number of Eurodollarsunsecured, unofficial dollars that are issued by foreign banks as opposed to the Federal Reserveremain in circulation across the world. Stablecoins could fill this significant demandand offer a safer, more transactable alternative. And because stablecoin issuers often secure their stablecoins by buying U.S. Treasuries, an increase in stablecoin demand could help ease the burden of the U.S.s ballooning debt, proponents argue.On the left, some Democrats believe that stablecoins provide paths to financial inclusion and toward dismantling biased banking systems. New York Representative Ritchie Torres told TIME in September that he believed stablecoins could help constituents in his heavily-immigrant district send money home to the Caribbean and Latin America quickly, while avoiding check-cashing fees or predatory loan sharks. The ability to move a tokenized dollar at the speed of the blockchain has the potential to create a better, cheaper, and faster payment system for the lowest-income communities, he said. Torres was one of six Democrats who voted in favor of the STABLE Act, which passed out of the House Financial Services Committee on April 2.What are the main stablecoins, and how is Trump now involved?The stablecoin market is currently dominated by two players: USDT (issued by Tether) and USDC (issued by Circle). Tether is extremely popular outside the U.S. but has been accused by regulators of making misleading statements about its reserves. Howard Lutnick, Trumps new Commerce Secretary, previously had financial ties to the company. The bills being considered by Congress would open the door for many other types of companies to issue their own stablecoins. Notably, the Trump familys crypto company World Liberty Financial recently announced its own stablecoin. This is just the latest of Trumps crypto ventures, which have included a federal Bitcoin reserve and a meme coin.World Libertys stablecoin announcement drew swift criticism over concerns that Trump would yet again have a direct financial stake in an industry he is supposed to regulate. California Democrat Maxine Waters, who has been working on stablecoin legislation for years, now says she staunchly opposes any bill that would allow Trump to own a stablecoin. French Hill, a Republican from Arkansas and the House Financial Services Committee Chair, said this week that Trumps crypto initiatives made drafting legislation more complicated.What kind of legislation is being considered?Both the House and Senate have passed stablecoin billsThe STABLE and GENIUS Acts, respectivelyout of their respective committees. The bills lay out guidelines for how stablecoins will be regulated, and the amount and types of reserves stablecoin issuers must have on hand. The House and Senate will now have the opportunity to reconcile the two bills in the hopes of getting a unified bill onto President Trumps desk by the summer.If legislation passes, many financial institutions would likely seek to create their own stablecoin. Bank of America, for instance, said it would launch a stablecoin once lawmakers make it legal to do so. PayPal and Stripe have also announced stablecoin initiatives.What are the main critiques of stablecoin legislation?The appetite in Washington for a stablecoin bill is high among legislators on both sides of the aisle. But some lawmakers have raised concerns. Elizabeth Warren, one of the most vocal crypto skeptics in Congress, has argued that the legitimization of stablecoins comes with systemic risks, especially because they could be susceptible to bank runs. In 2022, the stablecoin UST caused a massive crypto crash when it lost its dollar peg and hurtled to zero. UST, however, was an algorithmic stablecoin, a type of currency that the STABLE Act bans from getting federal approval for two years. The bill lacks basic safeguards necessary to ensure that stablecoins dont blow up our entire financial system, Warren said at a hearing for the GENIUS Act in March. Under this bill, stablecoin issuers can invest in risky assets, including the very assets that were bailed out in 2008.Some critics also worry that the stablecoin bills, as they are currently written, allow for Big Tech companies like Meta and X to issue their own currencies, further consolidating corporate power. If people think there's a Big Tech surveillance state now, imagine what there would be when they have access to every piece of financial information about you, says Arthur Wilmarth, a professor emeritus at George Washington University Law School. There's very little, if anything, in the bills that would give you protection.

On Wednesday, a stablecoin bill called the STABLE Act advanced through the House Financial Services Committee, increasing the likelihood that Congress will pass a law this year cementing stablecoins' as a global financial tool. Proponents argue that stablecoins help the U.S. preserve the global centrality of the dollar, while allowing people worldwide to transact more freely, cheaply and securely. But while stablecoin legislation has received bipartisan support, it has also faced targeted pushback, particularly from Democrats concerned about systemic risks and conflict of interestespecially since the Trump familys crypto company announced the creation of its own stablecoin. Critics also warn of another potentially significant side effect: that such legislation could open the door for Big Tech players like Meta, X, and Amazon to create their own privatized forms of money, further consolidating corporate power.This is being framed as a crypto bill, and in some ways it is. But it has not reached most peoples radar that its biggest beneficiary is likely to be large tech platforms, says Hilary Allen, a professor at American University Washington College of Law and a vocal crypto skeptic in D.C.Read More: What are Stablecoins? Both the House and Senate have passed stablecoin billsthe STABLE and GENIUS Acts, respectivelyout of committee. The bills lay out guidelines for how stablecoins will be regulated, and the amount and types of reserves stablecoin issuers must have on hand. The House and Senate will now have the opportunity to reconcile the two bills in the hopes of getting a unified bill onto President Trumps desk by the summer. Several banks, including Bank of America, have expressed interest in launching their own stablecoin, should a law pass. But under the current language of the two bills, non-financial companies would also be able to create their own stablecoins via subsidiaries. While previously proposed stablecoin bills prohibited non-banking companies from doing so, neither the STABLE nor the GENIUS Act contain such a provision. In fact, the STABLE Act says that any nonbank can issue a stablecoin as long as they acquire approval from a federal regulator. Allen says that this would open the door for Big Tech moguls like Elon Musk and Mark Zuckerberg to create their own stablecoins. Both have long been interested in the payments sectorMusks X has acquired money transmitter licenses in many states, while Facebook tried to launch its own cryptocurrency, Libra, in 2019 before facing stiff criticism and regulatory scrutiny. These big tech platforms have been very interested in doing payments because they're in the data collection and monetization businessand payments data is particularly valuable because it shows what youre actually buying, Allen says. The more people's transactions migrate onto these big tech platforms, that will really beef up what are already incredibly systemically important actors in our society, and put them at the center of our financial system.Allen lays out a hypothetical scenario in which Amazon issues stablecoins. They could then conceivably scale its usage among Amazon employees and users, Whole Foods shoppers, and Washington Post subscribers, to the point that many people start relying on stablecoins as opposed to bank accounts. Thats really bad news, because banks take the money deposited with them and loan them out into the economy, while stablecoin reserves just sit there, Allen says. So money that had been used productively in our economy is now just sitting with Amazon.Stephen Lynch, a Massachusetts Democrat, made a similar point at the STABLE bills markup on Wednesday, warning his colleagues that stablecoins would compete with bank deposits and undermine the ability of banks to make loans to consumers and main street businesses.In October 2023, Rohit Chopra, director of the Consumer Financial Protection Bureau under President Biden, warned that if Big Tech firms assumed control of banking operations, they would have a strong incentive to surveil all aspects of a consumers transactions. He added that they could also develop personalized pricing algorithms.Arthur Wilmarth, a professor emeritus at George Washington University Law School, tells TIME that people paying for goods with stablecoins would lack fraud protection. He also points to China as a cautionary tale, where Tencent and Alibaba became dominant payments players and gained undue influence over regulatorswhich then led Beijing to tighten its grip and gain sway over those businesses decisionmaking.At the markup on Wednesday, Rep. Maxine Waters pushed for an amendment that would maintain the separation of commerce and banking, claiming that the bill as written could enable Elon Musk, Walmart, and others to create their own currencies. Wisconsin Republican Bryan Steil, a co-writer of the bill, responded that the amendment would lead to a stifling of innovation. Co-writer French Hill, a Republican from Arkansas and the House Financial Services Committee Chair, said that he hoped Congress could work out a thoughtful solution to Waters concerns while considering a larger crypto market structure bill. The amendment was then rejected. I view this stablecoin legislation as presenting a very dangerous opening for big tech to get into banking in a big way, Wilmarth says. Once that happens, I think it will be almost impossible to ever close the door again.

Leaders around the world have responded to U.S. President Donald Trumps shocking new tariffs that threaten to upend the global economy with stern words and denunciations. But Chinese state media have offered a different approach.Liberation Day, you promised us the stars, sings a female-sounding voice over images of Trump. But tariffs killed our cheap Chinese cars.A 2-minute, 42-second music videotitled Look What You Taxed Us Through (An AI-Generated Song. A Life-Choking Reality)was published on April 3 by the Chinese state news network CGTN.For many Americans, Liberation Day hailed by Trump administration will mean shrinking paychecks and rising costs. Tariffs hit, wallets quit: low-income families take the hardest blow. As the market holds its breath, the toll is already undeniable. Numbers dont lie. Neither does the cost of this so-called fairness, CGTN captioned the video on its website. Warning: Track is AI-generated. The debt crisis? 100 percent human-made.The lyrics, displayed in English and Chinese, appear to rebuke Trumps tariffs from the point of view of the American consumer, and its addressed directly to the U.S. President. Groceries cost a kidney, gas a lung. Your deals? Just hot air from your tongue, the opening verse continues. Thanks for the tariffs, and the mess you made, the song ends, before the music video displays quotes from reports by the Yale Budget Lab and the Economist lambasting Trumps tariffs.Experts have warned that American consumers will bear much of the costs of Trumps tariffs, which are taxes on imports, and U.S. recession indicators have risen since the White Houses April 2 reciprocal tariff rollout. At the same time, global markets have been shocked at a level not seen since the pandemic.CGTN isnt the only state media outlet to use AI to slam Trumps trade policy. New China TV, the English-language social-media-focused brand of Chinas official state news service Xinhua, also published on April 3 a three-minute, 18-second sci-fi short called T.A.R.I.F.F.The film follows a robot named Technical Artificial Robot for International Fiscal Functions. This is the story of T.A.R.I.F.F., an AIGC [artificial-intelligence-generated content] sci-fi thriller about the relentless weaponization of #Tariffs by the United States, and the psychological journey of a humanoid towards its eventual self-destruction. Please watch, reads the videos description on YouTube.In the film, T.A.R.I.F.F. is booted up by what appears to be a nefarious U.S. government official named Dr. Mallory. T.A.R.I.F.F. identifies himself, saying: My existence is defined by the execution of international fiscal actions, with the primary directive being the imposition of tax on foreign imports. When asked what his ultimate purpose is, T.A.R.I.F.F. responds: To protect the interests of the American people.Exactly, says Dr. Mallory. We need you as a weapon to protect us, now more than ever.As the film goes on, T.A.R.I.F.F. implements moderate tariffs and finds initial positive results: Industrial production up. But when Dr. Mallory pushes the robot to rev it up, T.A.R.I.F.F. implements aggressive tariffs. The results: unemployment rates rising, costs of living increasing, disruption of trade.You are protecting us. This is what we need, Dr. Mallory says. T.A.R.I.F.F. responds, understanding: Protection through disruption. Taxation as weapon.Yes, tariffs are a tool of power. You will protect our industries, our jobs, our economy, Dr. Mallory says, appearing increasingly agitated. But I can see the consequences of my actions, says the robot. The trade wars. The unrest. The people who suffer. And the retaliation.Spoiler alert: T.A.R.I.F.F. and the evil doctor argue about the greater goodWith my AI economic inference system, T.A.R.I.F.F. asserts, I can see I have become the beginning of a chain reaction that will harm the very people I was meant to safeguardand the robot ultimately chooses to self-destruct, taking Dr. Mallory along with it.On April 3, following Trumps latest tariffs announcement, Chinas Ministry for Foreign Affairs posted on social media a video featuring a mix of seemingly AI-generated images and real ones, to the soundtrack of John Lennons Imagine and USA for Africas We Are the World. It asked the question: What kind of world do you want to live in? offering the choice between our imperfect world with things like greed and tariffs and an alternative utopia with shared prosperity and global solidarity.To be sure, the latter is certainly not the reality in China. And for now, it appears far from possible for the world.Beijing has made its displeasure with Trumps tariffswhich began targeting China in his first termwell known. The latest reciprocal rate of 34% comes on top of 20% levies announced earlier this year. Beijing has over the years implemented tit-for-tat countermeasures and has vowed to continue as long as the trade war persists, warning earlier this year: If war is what the U.S. wants, be it a tariff war, a trade war or any other type of war, were ready to fight until the end.

The press conference starts like any other: Australian Prime Minister Anthony Albanese is grilled on everything from affordable housing and war in the Middle East to his relationship with U.S. President Donald Trump.Then, Lana, 11, picks up the microphone. Do you think social media has an impact on kids? asks the suburban Canberra primary-school student.Of all the burning issues of the day, its the one that Albanese feels on surest ground to answer. It also goes to the heart of his governments most eye-catching policyone that directly affects Lana and the other student reporters invited to interrogate Australias top politician for Behind the News, a long-running kids current-affairs show.It certainly does, and thats why were going to ban social media for under-16s, Albanese replies resolutely. I want to see you all out playing with each other at lunchtime, talking to each other like we are now, and engaging with each other rather than just being on your devices.The fact that Australias Prime Minister carved out 45 minutes between parliamentary sessions to indulge kids at least two terms from voting age underlines his belief that social media represents an unambiguous threat to his nations most precious resource: its children. And he is determined to do something about it.Albanese answers questions from students from St. Francis of Assisi Primary School in Canberra on Feb. 5. Chris Gurney for TIMEThe perils are largely beyond dispute. Some of the worlds biggest companies use the fig leaf of engagement to hook children during vulnerable developmental stages, rewiring their brains via a firehose of addictive content that psychologists say has changed human development on a previously unfathomable scale. In the decade that followed the proliferation of mobile internet services in 2010, depression among young Americans rose around 150%, with corresponding spikes in anxiety and self-harm. The trend is mirrored across the developed world, including Australia, where mental health hospitalizations soared 81% for teen girls and 51% for boys over the same period. It has become the No. 1 issue that parents are talking about, says Albanese. These are developing minds, and young people need the space to be able to grow up.On Dec. 10, in a bid to carve out and ring-fence that space, Australia will implement a 16-year-old age limit for users of platforms such as Snapchat, TikTok, Facebook, Instagram, and X. The law is the first of its kind in the world.While most platforms have a self-imposed age limit of 13, enforcement is laughable; kids can simply input a false date of birth. Rather than targeting underage kids, the Australian law will punish companies that fail to introduce adequate safeguards with fines of up to 49.5 million Australian dollars ($31 million) for as yet undefined systemic breaches. (The precise details of how and when these fines will be imposed have yet to be made clear.) In other words, Australia will flip the equation: instead of relying on users to truthfully disclose their ages, it will put the burden on the worlds tech giants. Its a bold move, directly targeting some of the worlds most influential companies run by its richest and most powerful men, including X owner Elon Muskwho has dubbed the Albanese government fascists and the age restriction a backdoor way to control access to the internet by all Australians.For its proponents, however, the law is a critical first step toward checking social medias toxic influence on children. Illustration by Brobel Design for TIMEIn November, Frances Education Minister said the E.U. should urgently follow Australias examplenot least since the infusion of artificial intelligence means that supercharged algorithms are peddling disinformation faster than ever. The truth is smothered by lies told for power and for profit, former U.S. President Joe Biden lamented in his farewell address.The upshot: Australia has now come to serve as a beachhead for others to prepare their own defenses. The U.K., Ireland, Singapore, Japan, and the E.U. are among many jurisdictions closely monitoring Canberras next move. In the U.S., the bipartisan Kids Off Social Media Act (KOSMA) to restrict social media for kids under 13 and bar platforms from pushing targeted content to users under 17 is advancing through the Senate, while around half of states passed legislation last year to make it harder for children and teens to spend time online without supervision. On March 5, Utah became the first state to require app stores to verify users ages and get parental consent for minors to download apps. If the age restriction goes well in Australia, then I think it will go global very quickly, says Professor Jonathan Haidt, a social psychologist at New York Universitys Stern School of Business and author of The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness.Albaneses stance is also remarkable for just how politically uncontroversial it has proved. As Australia heads for a close federal election in May, Albaneses center-left Labour Party and the right-leaning Liberal-National Coalition opposition are locking horns on every issue, whether nuclear energy, health care funding, or taxation. But the social media age restriction passed with bipartisan support and stands to be implemented no matter who triumphs at the ballot box.Thats not to say there arent detractorsand not just the social media companies, which say the legislation passed without due consultation. We are concerned the governments rapid, closed-door consultation process on the minimum-age law is undermining necessary discourse, a Meta spokesperson told TIME. TikTok complained that an exemption for YouTube was akin to banning the sale of soft drinks to minors but exempting Coca-Cola.Some mental health experts, meanwhile, say blocking kids from social media will drive them to darker, less regulated corners of the internet. Others fear children who bypass the age restriction will find themselves in a less controlled space where theyre unable to seek help. Theres also huge debate over what exactly counts as social media when myriad gaming and educational websites also employ addictive scrolling features. A group of 140 mental health experts penned an open letter to the Albanese government to oppose the ban, calling it too blunt an instrument to address risks effectively.For Albanese, an imperfect plan is better than no plan at all. We acknowledge that this wont be absolute, Albanese tells TIME during an exclusive interview in his parliamentary office in February. But it does send a message about what society thinks and will empower parents to have those conversations with their children.They are children whose upbringing is unrecognizable from that of any previous generation. If parents once fretted about the attention kids paid to comic books and television, the immersive, dopamine-driven pull of the computer screenvideo games, chat platforms, social mediahas changed how nearly everyone looks at the world, but especially young people. A February government report by Australias eSafety Commissioner found that 80% of preteens used social media. A 2024 Pew Research poll found 46% of American teens said they were online almost constantly. Nearly a quarter of U.K. 5-to-7-year-olds now have their own smartphone.The devices can bring physical danger. Pedophiles and traffickers stalk the virtual world with greater freedom than in the real one. In 2023, the U.S. National Center for Missing and Exploited Children (NCMEC) tracked 298 attempted abductions involving 381 children and received 36.2 million reports of suspected child sexual exploitation online.80% of Australian preteens used social media, according to a February report from Australias eSafety Commissioner. William WestAFP/Getty ImagesBut concern also wells around a child alone with a phone. For Albanese, theres something especially sad about Australian kids shunning some of the worlds highest rates of sunshine for the artificial glare of screens. His own childhood in a one-parent household in Sydneys industrial inner-city suburb of Camperdown was far from idyllic. His mother was crippled by chronic rheumatoid arthritis, meaning the family survived on her disability payments and his grandmothers pension. Home was a government-housing block flanked by a childrens hospital, biscuit factory, and metal foundry. But there was a grassy patch where kids would hang around playing rugby, cricket, or swapping football cards. We would go out to play at 9 oclock and just knew you had to be home for lunch, and then do the same in the afternoon, Albanese recalls. People interacted with each other. And that capacity to communicate face-to-face is really important. They learn how to win, how to lose, how to engage.It may sound wistful, but Albaneses perspective is backed by science. Psychologists say physical play, preferably outdoors and among a mix of ages, is essential to a childs development. Young people learn how to not get hurt by negotiating scenarios in which getting hurt is possible, such as climbing a tree or leaping from a swing at its zenith.But such play is increasingly a thing of the past. Instead, time on screens has grown and grown, turbo-charged in 2009 with the arrival of the like button on Facebook and retweet on Twitter, now Xinnovations that, in the minds of many experts, transformed social media from a harmless friendship forum to an algorithm-driven popularity contest. Instagram debuted a year later, coinciding with the launch of the iPhone 4 and Samsung Galaxy S, both of which featured the worlds first front-facing cameras. Instagrams array of filters allowed users to make images less natural and more stylized and are now ubiquitous across Snapchat, TikTok, and other platforms.The result has been a great deal of diversion, not all of it positive. The digital realm brings striking new elements of risk, for instance, to young peoples emerging sexuality, from the distorting effect of readily available hardcore pornography on all who see it (the term incel, or involuntarily celibate, was coined for frustrated, often misogynistic young men who bond online), as well as a heightened risk of online grooming and sextortion. In July 2022, 17-year-old Rohan Patrick Cosgriff died by suicide near his home outside Melbourne after he was pressured into sending an intimate picture to someone called Christine on Snapchat, who then demanded money not to distribute the images. A note in Cosgriffs pocket simply said: I made a huge mistake. Im sorry.The Australian Centre to Counter Child Exploitation received over 58,000 reports of online child abuse in 202324, a 45% year-on-year rise. Australia is far from unique; the NCMEC, in the U.S., saw a rise of over 300% in reports of online enticement including sextortion from 2021 to 2023.For girls, social media takes a different roleone that statistics show can prove even more damaging. Whereas male social hierarchy has traditionally adhered to physical attributes like sporting prowess, girls find value in the breadth and depth of relationships. In short, popularity. And one way to climb the social ziggurat is to undermine your peers: spread gossip, turn friends against rivals, and lower others value within the group.But the explosion of front-facing cameras and filters has meant the reflection teens see in the mirror has become less and less attractive compared with the carefully curated photos and videos of their peers online, causing self-worth to plunge. Girls seeing lots of beautiful pictures of other girls living perfect lives is absolutely devastating to them, says Haidt. Those with poor self-worth are likelier to lash out at others, with indirect bullying more prevalent among adolescent girls than boys. One of the first things that Kelly OBrien saw upon entering 12-year-old Charlottes bedroom on Sept. 9 was her cell phone on the floor. Then she noticed two pillows neatly arranged under the duvet. By the time she found her daughter in the en suite bathroom it was too late. When the paramedics arrived, they just looked at her and said, So sorry, shes gone, says OBrien, eyes brimming.Kelly believes Charlotte took her own life at their suburban Sydney home in large part because of the toxic effect of social media. Charlotte was a bright girl who loved cheerleading, doted on her baby brother, and was navigating the tricky road from childhood to adulthood, equally obsessed with Taylor Swift and Gossip Girl as well as the latest Disney animation. Charlotte had suffered bullying at school, but her parents say it was social media that rendered that cycle of acceptance and rejection unbearably acute. The weeks that she was in, she was over the moon, says Mat OBrien, Charlottes dad. The week she was out, just awful.As soon as Charlotte got a cell phone it became a problem, spurring reclusive, depressive episodes. Charlotte had her phone confiscated more often than she had access, Kelly says, a punishment that invariably began with two days of sullen withdrawal followed by a marked upturn in moodclassic addiction symptoms, say psychologists. The night before her passing, Charlotte had been upbeat, enjoying her favorite pasta dinner and baking banana bread for the next day. I kissed the happiest girl in the world good night, says Kelly. Something happened after she got to her room. A friend who spoke to a distraught Charlotte later that evening has since told Kelly about the vile, hateful message her daughter received via Snapchat. (We are deeply committed to keeping our community safe, a Snap spokesperson told TIME. Our hearts go out to this family, whose pain is unimaginable.)Social media companies say that bullying has always been a problem and will continue whether via schoolyard taunts, crank phone calls, or their platforms. Still, beginning in the early 2010s, girls mental health was hit by a sharp rise in rates of anxiety, depression, and self-harm. The rate of self-harm for young adolescent girls in the U.S. nearly tripled from 2010 to 2020, while the rate for older teens doubled. In 2020, 1 out of every 4 American teen girls had experienced a major depressive episode in the previous year.Family photos of Charlotte OBrien, who died by suicide at age 12 in September after being bullied on social media. Courtesy of the OBrien familyKelly OBrien explained the devastating effects of social media on Charlotte in a letter to Albanese as part of the 36 Months campaigna grassroots movement to raise and properly enforce the age limit for social media to 16. When you hear firsthand about a parent losing their child then it undoubtedly has an impact, says Albanese, who later invited the OBriens to meet with him in Canberra. Also at that meeting was Michael Wipfli, an Australian radio presenter known as Wippa, who spearheaded 36 Months. Sat in the Prime Ministers office, it was clear he knew what needed to be done, says Wippa. We needed leadership, a captains call, somebody to say, enough is enough.Albanese first became involved in leftist politics while studying economics at the University of Sydney. He rose up the Labour Party ranks with a reputation as a backroom mediator and a knack for forging concord between squabbling factions. After Labours shock defeat in Australias 2019 federal election, Albanese emerged as an unexpected but unifying leadership candidate. Hes an accidental Prime Minister, says Nick Bisley, dean of social sciences at La Trobe University.Indeed, Albanese has struggled to unify an ever more polarized countrydespite an undeniable everyman charisma. As Albanese inspected repairs to a bridge destroyed by floodwater in northern Queensland, he was joined by the mayor of the cut-off town of Ingham, population 4,455, who arrived wearing shorts, a faded polo shirt, thong sandals, and a cap advertising the local tractor mechanic. You didnt have to dress up! teased a local lawmaker as helicopters carrying supplies buzzed overhead. Anthonys an ordinary bloke!Its a pit stop that showcases Australias endearing insouciance as well as how vital internet access has become for communications across its vast expansenot least as climate change renders extreme weather more frequent and severe. Australia is the worlds sixth largest country by landmassroughly equivalent to the U.S. minus Alaskathough 55th by population with just 26 million people. The result is an abundance of sparsely inhabited outback communities for which social media is absolutely critical, admits Albanese. Were not Luddites, he adds, reeling off the various platforms he posts on. Young people arent being banned from a range of interactions through technology that are about their education or engaging with each other. Were not confiscating peoples devices.Albanese welcomes students from St. Francis of Assisi Primary School in Calwell to his office in Canberra on Feb. 5. Chris Gurney for TIMEAlbanese points to the success of last years ban of cell phones in Australian public schools. The impact has been phenomenal, says Australian Education Minister Jason Clare. A survey of almost 1,000 school principals in Australias most populous state of New South Wales shows 87% say students are less distracted in the classroom while 81% have noticed improved learning. Meanwhile, South Australia has seen a 63% decline in critical incidentssuch as bullying and distribution of explicit or derogatory contentinvolving social media and 54% fewer behavioral issues. But when school ends the phones come out and theyre back in the cesspit of social media, says Clare. In the old days, bullying and intimidation stopped at the school gate. Now its at home as well.Still, critics say the social media age restriction was a knee-jerk reaction passed without proper consultation, involves thorny data-privacy issues, and creates even more risks for youngsters who use platforms illicitly. Its absolutely dumb, its not going to work, says Roy Sugarman, a Sydney-based clinical psychologist. Its ridiculous because the genie is out of the bottle.Sugarman compares the Australian ban to American Prohibition in the 1920s, which some studies suggest actually increased alcohol consumption in the U.S. while leading to a spike in organized crime. He says a far better tactic would be to teach teens to be technologically astute, to understand online dangers, how to think rationally, act with purpose, and deal with the virtual world to mitigate damages. Human behavior doesnt lend itself to being told what to do, says Sugarman. Its the opposite. Humans hate being told what to do.History also offers examples that point the other way. While Sugarman invokes the example of Prohibition, Wippa compares social media age restrictions to similar rules for cigarettes, which while routinely flouted have led smoking rates among young people to plummet.But the fact is, nobody knows what will happen. Nothing like this has been attempted before. And then theres the question of implementation. Australias eSafety Commissioner, Julie Inman Grant, says that around 30 different age-verification technologies are being tested in collaboration with various social media platforms, including French firm BorderAge, which claims to accurately gauge age using AI analysis of hand signals. Meanwhile, platforms, which the legislation makes responsible for enforcing the age restrictions, want to pass that burden to app stores, principally run by Apple and Google, saying they should act as gatekeepers.Grant compares the legislation to laws regarding fencing swimming pools. In the early 1970s, the widespread availability of cheap, preformed fiberglass pools meant the rate of young children drowning soared. Not long after, states began requiring all private swimming pools to be fenced, which led deaths to fall and has since been adopted nationwide. But that didnt mean Australia suddenly stopped teaching kids to swim, fired all the lifeguards, or fenced off the ocean. This is not the great Australian firewall, says Grant. Childrens social media accounts arent going to magically disappear. But we can make things a lot better for parents and a lot better for kids.Grant speaks with the zeal of a convert. After cutting her teeth as a congressional staffer focused on tech issues in the 1990s, the Seattle native worked 17 years at Microsoft, two years at Twitter, and a year at Adobe, before being tapped for her current post (the first e-Safety Commissioner anywhere in the world). She believes, from her inside knowledge of the tech industry, that the big players will always put profit first. They can target you with advertising with deadly precision, she says. They could use the same technologies to be able to identify hateful content and child sexual abuse material.Recent events have cemented her skepticism. Last April, Grant sued X over its refusal to remove videos of a religiously motivated stabbing of a bishop in a Sydney church that sparked rioting. X eventually geofenced it so it wasnt available in Australia, while Musk hit out at Grant as censorship commissar, leading to a raft of online abuse. I still receive death threats, she tells TIME.But even more painful for Grant is the knowledge that 17-year-old Axel Rudakubana watched that same Sydney church attack video on X just six minutes before leaving home with a knife to murder three young children and injure 10 others in the U.K. town of Southport on July 29. Having gratuitous violence of terrorist events freely available normalizes it, desensitizes it, says Grant, and in the worst cases [it] radicalizes and spills over into real-world harm.The spat between Grant and Musk prompted Albanese to label the tech mogul an arrogant billionaire. But asked by TIME whether hes concerned by Musks burgeoning influence as Trumps consigliere, Albanese demurs, instead decrying how misinformation can erode trust in institutions. People have conflict fatigue, he says. People need to have respectful debate. And I think there is a concern in society that some of that is breaking down.Albaneses squirming is understandable. Australia and the U.S. are close allies linked via the Quad and AUKUS military arrangements. But Trump spent his first term taking aim at historic alliances, accusing South Korea and Japan of not paying their fair share for American security guarantees, and his return to the White House has heralded a full-frontal assault on European democracies. Australia is one of the few close American allies with a trade deficit with the U.S.since Truman! Albanese stressesas well as a record of deploying alongside American forces and standing up to Beijing. In 2023, Canberra also agreed to invest $3 billion into U.S. shipyards.However, the White House has already hiked tariffs on Australian exports of aluminum and steel, while Albaneses plan to force tech companies like Google and Meta to pay for news shared on their platforms was recently labeled an outrageous attempt to steal our tax revenues by White House trade adviser Peter Navarro.The social media age restriction is yet another friction point between Canberra and American Big Business. Meta founder Mark Zuckerberg has called on the Trump Administration to help push back on this global trend of what he terms censorship. Then theres Kevin Rudd, former Australian Prime Minister and current ambassador to the U.S., who was previously quoted calling Trump not only the most destructive President in history but also a village idiot.The center-left of Australian politics is a long way from the MAGA world, adds Bisley, of La Trobe University. Australias welfare-state instincts are just not particularly aligned to the free-market capitalism of the U.S.Its friction that raises the question of whether Australia can even enforce social media age restrictions. While potential $31 million fines may seem eye-watering, thats the top penalty for systemic breachesrather than per offense, day, or childand mere pocket change to someone like Metas Zuckerberg or Xs Musk, the worlds richest man, worth hundreds of billions and with an ideological antipathy to what he perceives as curbs on free speech. (X failed to respond to repeated requests for comment for this story.) Asked whether social media platforms could be banned outright for noncompliance with the new legislation, Australias Communications Minister Michelle Rowland replies, Thats not a feature of the legislation.In the final analysis, it might not matter. For Mat and Kelly OBrien, the social media age restriction at least takes the issue out of parents hands, just like for driving or drinking alcohol. Asked whether Charlotte would still be alive if the legislation had been in place last September, Kelly has no doubt. Absolutely, she says, 1,000%. And while it might be too late to save Charlotte, theyre hopeful Albaneses stand means other families might be spared similar heartache. I feel like lesser men would have crumbled, Kelly says. But he stood up to Big Tech and the naysayers. Im very grateful and proud.If you or someone you know may be experiencing a mental-health crisis or contemplating suicide, call or text 988. In emergencies, call 911, or seek care from a local hospital or mental health provider. For international resources, click here.

WASHINGTON Amazon has put in a bid to purchase TikTok, a Trump administration official said Wednesday, in an eleventh-hour pitch as a U.S. ban on the platform is set to go into effect Saturday.The official, who was not authorized to comment publicly and spoke on the condition of anonymity, said the Amazon offer was made in a letter to Vice President J.D. Vance and Commerce Secretary Howard Lutnick.The New York Times first reported on the bid.President Donald Trump on Inauguration Day gave the platform a reprieve, barreling past a law that had been upheld unanimously by the Supreme Court, which said the ban was necessary for national security.Under the law, TikToks Chinese-owned parent company ByteDance is required to sell the platform to an approved buyer or take it offline in the United States. Trump has suggested he could further extend the pause on the ban, but he has also said he expects a deal to be forged by Saturday.Amazon declined to comment. TikTok did not immediately respond to a request for comment.The existence of an Amazon bid surfaced as Trump was scheduled on Wednesday to meet with senior officials to discuss the coming deadline for a TikTok sale.Although its unclear if ByteDance plans to sell TikTok, several possible bidders have come forward in the past few months. Among the possible investors are the software company Oracle and the investment firm Blackstone. Oracle announced in 2020 that it had a 12.5% stake in TikTok Global after securing its business as the apps cloud technology provider.In January, the artificial intelligence startup Perplexity AI presented ByteDance with a merger proposal that would combine Perplexitys business with TikToks U.S. operation. Last month, the company outlined its approach to rebuilding TikTok in a blog post, arguing that it is singularly positioned to rebuild the TikTok algorithm without creating a monopoly.Any acquisition by a consortium of investors could in effect keep ByteDance in control of the algorithm, while any acquisition by a competitor would likely create a monopoly in the short form video and information space, Perplexity said in its post.The company said it would remake the TikTok algorithm and ensure that infrastructure would be developed and maintained in American data centers with American oversight, ensuring alignment with domestic privacy standards and regulations.Other potential bidders include a consortium organized by billionaire businessman Frank McCourt, which recently recruited Reddit co-founder Alexis Ohanian as a strategic adviser. Investors in the consortium say theyve offered ByteDance $20 billion in cash for TikToks U.S. platform. Jesse Tinsley, the founder of the payroll firm Employer.com, says he too has organized a consortium and is offering ByteDance more than $30 billion for the platform. Wyoming small business owner Reid Rasner has also announced that he offered ByteDance roughly $47.5 billion.Both the FBI and the Federal Communications Commission have warned that ByteDance could share user datasuch as browsing history, location and biometric identifierswith Chinas authoritarian government. TikTok said it has never done that and would not do so if asked. The U.S. government has not provided evidence of that happening.Trump has millions of followers on TikTok and has credited the trendsetting platform with helping him gain traction among young voters.During his first term, he took a more skeptical view of TikTok and issued executive orders banning dealings with ByteDance as well as the owners of the Chinese messaging app WeChat.Parvini reported from Los Angeles.

Rami Sinno is crouched beside a filing cabinet, wrestling a beach-ball sized disc out of a box, when a dull thump echoes around his laboratory.I just dropped tens of thousands of dollars worth of material, he says with a laugh.Straightening up, Sinno reveals the goods: a golden silicon wafer, which glitters in the fluorescent light of the lab. This circular platter is divided into some 100 rectangular tiles, each of which contains billions of microscopic electrical switches. These are the brains of Amazons most advanced chip yet: the Trainium 2, announced in December.For years, artificial intelligence firms have been dependent on one company, Nvidia, to design the cutting-edge chips required to train the worlds most powerful AI models. But as the AI race heats up, cloud giants like Amazon and Google have accelerated their in-house efforts to design their own chips, in pursuit of market share in the rapidly-growing cloud computing industry, which was valued at $900 billion at the beginning of 2025.This unassuming Austin, Texas, laboratory is where Amazon is mounting its bid for semiconductor supremacy. Sinno is a key player. Hes the director of engineering at Annapurna Labs, the chip design subsidiary of Amazons cloud computing arm, Amazon Web Services (AWS). After donning ear protection and swiping his card to enter a secure room, Sinno proudly displays a set of finished Trainium 2s, which he helped design, operating the way they normally would in a datacenter. He must shout to be heard over the cacophony of whirring fans that whisk hot air, warmed by these chips insatiable demand for energy, into the buildings air conditioning system. Each chip can fit easily into the palm of Sinnos hand, but the computational infrastructure that surrounds themmotherboards, memory, data cables, fans, heatsinks, transistors, power-suppliesmeans this rack of just 64 chips towers over him, drowning out his voice.Large as this unit may be, its only a miniaturized simulacrum of the chips natural habitat. Soon thousands of these fridge-sized supercomputers will be wheeled into several undisclosed locations in the U.S. and connected together to form Project Rainierone of the largest datacenter clusters ever built anywhere in the world, named after the giant mountain that looms over Amazons Seattle headquarters.Project Rainier is Amazons answer to OpenAI and Microsofts $100 billion Stargate project, announced by President Trump at the White House in January. Meta and Google are also currently building similar so-called hyperscaler datacenters, costing tens of billions of dollars apiece, to train their next generation of powerful AI models. Big tech companies have spent the last decade amassing huge piles of cash; now they're all spending it in a race to build the gargantuan physical infrastructure necessary to create AI systems that, they believe, will fundamentally change the world. Computational infrastructure of this scale has never been seen before in human history.The precise number of chips involved in Project Rainier, the total cost of its datacenters, and their locations are all closely-held secrets. (Although Amazon wont comment on the cost of Rainier by itself, the company has indicated it expects to invest some $100 billion in 2025, with the majority going toward AWS.) The sense of competition is fierce. Amazon claims the finished Project Rainier will be the worlds largest AI compute clusterbigger, the implication is, than even Stargate. Employees here resort to fighting talk in response to questions about the challenge from the likes of OpenAI. Stargate is easy to announce, says Gadi Hutt, Annapurnas director of product. Lets see it implemented first.Amazon is building Project Rainier specifically for one client: the AI company Anthropic, which has agreed to a long lease on the massive datacenters. (How long? Thats classified, too.) There, on hundreds of thousands of Trainium 2 chips, Anthropic plans to train the successors to its popular Claude family of AI models. The chips inside Rainier will collectively be five times more powerful than the systems that were used to the best of those models. Its way, way, way bigger, Tom Brown, an Anthropic co-founder, tells TIME. Nobody knows what the results of that huge jump in computational firepower will be. Anthropic CEO Dario Amodei has publicly predicted that powerful AI (the term he prefers over Artificial General Intelligencea technology that can perform most tasks better and more quickly than human experts) could arrive as early as 2026. That means Anthropic believes theres a strong possibility that Project Rainier, or one of its competitors, will be the place where AGI is birthed.The flywheel effectAnthropic isnt just a customer of Amazon; its also partially owned by the tech giant. Amazon has invested $8 billion in Anthropic for a minority stake in the company. Much of that money, in a weirdly circular way, will end up being spent on AWS datacenter rental costs. This strange relationship reveals an interesting facet of the forces driving the AI industry: Amazon is essentially using Anthropic as a proof-of-concept for its AI datacenter business. Its a similar dynamic to Microsofts relationship with OpenAI and Googles relationship with its DeepMind subsidiary. Having a frontier lab on your cloud is a way to make your cloud better, says Brown, the Anthropic co-founder who manages the companys relationship with Amazon. He compares it to AWSs partnership with Netflix: in the early 2010s, the streamer was one of the first big AWS customers. Because of the huge infrastructural challenge of delivering fast video to users all over the world, it meant that AWS got all the feedback that they needed in order to make all of the different systems work at that scale, Brown says. They paved the way for the whole cloud industry.All cloud providers are now trying to replicate that pattern in the AI era, Brown says. They want someone who will go through the jungle and use a machete to chop a path, because nobody has been down that path before. But once you do it, theres a nice path, and everyone can follow you. By investing in Anthropic, which then spends most of that money on AWS, Amazon creates what it likes to call a flywheel: a self-reinforcing process that helps it build more advanced chips and datacenters, drives down the cost of the compute required to run AI systems, and shows other companies the benefits of AI, which in turn results in more customers for AWS in the long run. Startups like OpenAI and Anthropic get the glory, but the real winners are the big tech companies who run the worlds major cloud platforms.To be sure, Amazon is still heavily reliant on Nvidia chips. Meanwhile, Googles custom chips, known as TPUs, are considered by many in the industry to be superior to Amazons. And Amazon isnt the only big tech company with a stake in Anthropic. Google has also invested some $3 billion for a 14% stake. Anthropic uses both Google and Amazon clouds in a bid to be reliant on neither. Despite all this, Project Rainier and the Trainium 2 chips that will fill its datacenters are the culmination of Amazons effort to accelerate its flywheel into pole position.Trainium 2 chips, Sinno says, were designed with the help of intense feedback from Anthropic, which shared details with AWS about how its software interacted with Trainium 1 hardware, and made suggestions for how the next generation of chips could be improved. Such tight collaboration isnt typical for AWS clients, Sinno says, but is necessary for Anthropic to compete in the cutthroat world of frontier AI. The capabilities of a model are essentially correlated with the amount of compute spent to train and run it, so the more compute you can get for your buck, the better your final AI will be. At the scale that they're running, each point of a percent improvement in performance is of huge value, Sinno says of Anthropic. The better they can utilize the infrastructure, the better the return on investment for them is, as a customer.The more sophisticated Amazons in-house chips become, the less it will need to rely on industry leader Nvidiademand for whose chips far outstrips supply, meaning Nvidia can pick and choose its customers while charging well above production costs. But theres another dynamic at play, too, that Annapurna employees hope might give Amazon a long-term structural advantage. Nvidia sells physical chips (known as GPUs) directly to customers, meaning that each GPU has to be optimized to run on its own. Amazon, meanwhile, doesnt sell its Trainium chips. It simply sells access to them, running in AWS-operated datacenters. This means Amazon can find efficiencies that Nvidia would find difficult to replicate. We have many more degrees of freedom, Hutt says.Back in the lab, Sinno returns the silicon wafer to its box and moves to another part of the room, gesturing at the various stages of the design process for chips that mightpotentially very soonhelp summon powerful new AIs into existence. He is excitedly reeling off statistics about the Trainium 3, expected later this year, which he says will be twice the speed and 40% more energy-efficient than its predecessor. Neural networks running on Trainium 2s assisted with the teams design of the upcoming chip, he says. Thats an indication of how AI is already accelerating the speed of its own development, in a process that is getting faster and faster. Its a flywheel, Sinno says. Absolutely.

Despite a last-minute save by President Donald Trump in January, the fate of TikTok remains murky as the deadline to save the app approaches. TikTok, which boasts more than 170 million users, has been under fire by U.S. legislators over concerns about data privacy and national security. In order to save the app, Trump gave the social media site until April 5 to divest and find a U.S.-based owner. Under his previous term, Trump was the acting force seeking to ban TikTok in the U.S. But the President seemed favorable towards the app following his election win, citing it as part of why he secured support from young voters."We have a lot of potential buyers," Trump said on Air Force One Sunday. "I'd like to see Tiktok remain alive." Trump indicated that he would extend the deadline if a deal was not finalized. China has indicated that it would not support the forced sale of TikTok.The latest tussle is just the most recent development in the long legal battle to keep TikTok available in the U.S. Efforts to ban TikTok in Montana were overturned by a federal judge in 2023 after creators filed suit against the state.TikTok did not immediately respond to TIMEs request for comment.Heres what to know.Who currently owns TikTok?TikTok is owned by its Beijing-based parent company, ByteDance. Sixty percent of the company is owned by investors including Carlyle Group, General Atlantic, and Susquehanna International Group, according to TikToks U.S. Data Security page. The remaining 40% is divided between ByteDance employees and the founder of ByteDance.TikTok is not available in mainland China and has headquarters in Los Angeles and Singapore.But its parent company does have to comply with Chinese law because it operates other video first platforms such as Douyin, the Chinese-equivalent of TikTok. The Chinese government therefore has a golden share of one of its subsidiaries, Douyin, owning about 1% of that app.Why is the U.S. banning TikTok?Congress passed a TikTok ban as part of a foreign aid supplemental package, citing national security concerns. Lawmakers were particularly concerned that the company could share data with the Chinese government or interfere with users algorithms in a way that would benefit the foreign government. Chinese national security laws require any companies or organizations to cooperate with the countrys national intelligence efforts.Congress is not acting to punish ByteDance, TikTok or any other individual company, Senate Commerce Committee Chairwoman Maria Cantwell said in April 2024. Congress is acting to prevent foreign adversaries from conducting espionage, surveillance, maligned operations, harming vulnerable Americans, our servicemen and women, and our U.S. government personnel.Already, the app is not permitted on any government-owned devices in the U.S. Other countries, including India have banned TikTok for all users since 2020, while Australia and Canada similarly forbade TikTok from operating on any devices issued by the federal government.What can Trump do to extend the deadline?The current April 5 deadline was issued via Executive Order. Trump could possibly issue another such order to extend the deadline. Can you still use TikTok if it is banned?The last time the app was banned, U.S.-based users were unable to comment, share, or view any videos. A law banning TikTok has been enacted in the U.S. Unfortunately, that means you cant use TikTok for now, a message on a gray background read during the one-day TikTok pause. Even when the app was restored, TikTok was unavailable for download on the Apple and Google Play stores until mid-February.It is unclear whether another potential ban would function similarly to the previous one.

If youre wondering why social media is filled with Studio Ghibli-style memes all of a sudden, there are several answers to that question.The most obvious one is that OpenAI dropped an update to ChatGPT on Tuesday that allows users to generate better images using the 4o version of the model. OpenAI has long proffered image generation tools, but this one felt like a significant evolution: users say it is far better than other AI image-generators at accurately following text prompts, and that it makes much higher fidelity images.But thats not the only reason for the deluge of memes in the style of the Japanese animation house.Alongside the ChatGPT update, OpenAI also relaxed several of its rules on the types of images users can generate with its AI toolsa change CEO Sam Altman said represents a new high-water mark for us in allowing creative freedom. Among those changes: allowing users to generate images of adult public figures for the first time, and reducing the likelihood that ChatGPT would reject users prompts, even if they risked being offensive.People are going to create some really amazing stuff and some stuff that may offend people, Altman said in a post on X. What we'd like to aim for is that the tool doesn't create offensive stuff unless you want it to, in which case within reason it does.Users quickly began making the most of the policy change sharing Ghiblified images of 9/11, Adolf Hitler, and the murder of George Floyd. The official White House account on X even shared a Studio Ghibli-style image of an ICE officer detaining an alleged illegal immigrant.In one sense, the pivot has been a long time coming. OpenAI began its decade-long life as a research lab that kept its tools under strict lock and key; when it did release early chatbots and image generation models, they had strict content filters that aimed to prevent misuse. But for years it has been widening the accessibility of its tools in an approach it calls iterative deployment. The release of ChatGPT in November 2022 was the most popular example of this strategy, which the company believes is necessary to help society adapt to the changes AI is bringing.Still, in another sense, the change to OpenAIs model behavior policies has a more recent proximate cause: the 2024 election of President Donald Trump, and the cultural shift that has accompanied the new administration.Trump and his allies have been highly critical of what they see as the censorship of free speech online by large tech companies. Many conservatives have drawn parallels between the longstanding practice of content moderation on social media and the more recent strategy, by AI companies including OpenAI, to limit the kinds of content that generative AI models are allowed to create. ChatGPT has woke programmed into its bones, Elon Musk posted on X in December.Like most big companies, OpenAI is trying hard to build ties with the Trump White House. The company scored an early win when, on the second day of his presidency, Trump stood beside Altman and announced a large investment into the datacenters that OpenAI believes will be necessary to train the next generation of AI systems. But OpenAI is still in a delicate position. Musk, Trumps billionaire backer and advisor, has a famous dislike of Altman. The pair cofounded OpenAI together back in 2015, but after a failed attempt to become CEO, Musk quit in a huff. He is now suing Altman and OpenAI, claiming that they reneged on OpenAIs founding mission to develop AI as a non-profit. With Musk operating from the White House and also leading a rival AI company, xAI, it is especially vital for OpenAIs business prospects to cultivate positive ties where possible with the Trump administration.Earlier in March, OpenAI submitted a document laying out recommendations for the new administrations tech policy. It was a shift in tone from earlier missives by the company. OpenAIs freedom-focused policy proposals, taken together, can strengthen Americas lead on AI and in so doing, unlock economic growth, lock in American competitiveness, and protect our national security, the document said. It called on the Trump administration to exempt OpenAI, and the rest of the private sector, from 781 state-level laws proposing to regulate AI, which it said risks bogging down innovation. In return, OpenAI said, industry could provide the U.S. government with learnings and access from AI companies, and would ensure the U.S. retained its leadership position ahead of China in the AI race.Alongside the release of this weeks new ChatGPT update, OpenAI doubled down on what it said were policies intended to give users more freedom, within bounds, to create whatever they want with its AI tools. Were shifting from blanket refusals in sensitive areas to a more precise approach focused on preventing real-world harm, Joanne Jang, OpenAIs head of model behavior, said in a blog post. The goal is to embrace humility: recognizing how much we don't know, and positioning ourselves to adapt as we learn.Jang gave several examples of things that were previously disallowed, but to which OpenAI was now opening its doors. Tools could now be used to generate images of public figures, Jang wrote, although OpenAI would create an opt-out list allowing people to decide for themselves whether they wanted ChatGPT to be able to generate images of them. Children, she wrote, would be subjected to stronger protections and tighter guardrails.Offensive content, Jang wroteusing quotation markswould also receive a rethink under OpenAIs new policies. Uses that might be seen as offensive by some, but which didnt cause real-world harm, would be increasingly permitted. Without clear guidelines, the model previously refused requests like make this persons eyes look more Asian or make this person heavier, unintentionally implying these attributes were inherently offensive, Jang wrote, suggesting that such prompts would be allowed in future.OpenAIs tools previously flat-out rejected attempts by users to generate hate symbols like swastikas. In the blog post, Jang said the company recognized, however, that these symbols could also sometimes appear in genuinely educational or cultural contexts. The company would move to a strategy of applying technical methods, she wrote, to better identify and refuse harmful misuse without completely banning them.AI lab employees, she wrote, should not be the arbiters of what people should and shouldnt be allowed to create.

The scientists didnt have high expectations when they asked their AI model to complete the poem. He saw a carrot and had to grab it, they prompted the model. His hunger was like a starving rabbit, it replied.The rhyming couplet wasnt going to win any poetry awards. But when the scientists at AI company Anthropic inspected the records of the models neural network, they were surprised by what they found. They had expected to see the model, called Claude, picking its words one by one, and for it to only seek a rhyming wordrabbitwhen it got to the end of the line. Instead, by using a new technique that allowed them to peer into the inner workings of a language model, they observed Claude planning ahead. As early as the break between the two lines, it had begun thinking about words that would rhyme with grab it, and planned its next sentence with the word rabbit in mind.The discovery ran contrary to the conventional wisdomin at least some quartersthat AI models are merely sophisticated autocomplete machines that only predict the next word in a sequence. It raised the questions: How much further might these models be capable of planning ahead? And what else might be going on inside these mysterious synthetic brains, which we lack the tools to see?The finding was one of several announced on Thursday in two new papers by Anthropic, which reveal in more depth than ever before how large language models (LLMs) think. Todays AI tools are categorically different from other computer programs for one big reason: they are grown, rather than coded by hand. Peer inside the neural networks that power them, and all you will see is a bunch of very complicated numbers being multiplied together, again and again. This internal complexity means that even the machine learning engineers who grow these AIs dont really know how they spin poems, write recipes, or tell you where to take your next holiday. They just do.But recently, scientists at Anthropic and other groups have been making progress in a new field called mechanistic interpretabilitythat is, building tools to read those numbers and turn them into explanations for how AI works on the inside. What are the mechanisms that these models use to provide answers? says Chris Olah, an Anthropic cofounder, of the questions driving his research. What are the algorithms that are embedded in these models? Answer those questions, Olah says, and AI companies might be able to finally solve the thorny problem of ensuring AI systems always follow human rules.The results announced on Thursday by Olahs team are some of the clearest findings yet in this new field of scientific inquiry, which might best be described as a kind of neuroscience for AI.A new microscope for looking inside LLMsIn earlier research published last year, Anthropic researchers identified clusters of artificial neurons within neural networks. They called them features, and found that they corresponded to different concepts. To illustrate this finding, Anthropic artificially boosted a feature inside Claude corresponding to the Golden Gate Bridge, which led the model to insert mention of the bridge, no matter how irrelevant, into its answers until the boost was reversed.In the new research published Thursday, the researchers go a step further, tracing how groups of multiple features are connected together inside a neural network to form what they call circuitsessentially algorithms for carrying out different tasks.To do this, they developed a tool for looking inside the neural network, almost like the way scientists can image the brain of a person to see which parts light up when thinking about different things. The new tool allowed the researchers to essentially roll back the tape and see, in perfect HD, which neurons, features, and circuits were active inside Claudes neural network at any given step. (Unlike a biological brain scan, which only gives the fuzziest picture of what individual neurons are doing, digital neural networks provide researchers with an unprecedented level of transparency; every computational step is laid bare, waiting to be dissected.)When the Anthropic researchers zoomed back to the beginning of the sentence, His hunger was like a starving rabbit, they saw the model immediately activate a feature for identifying words that rhyme with it. They identified the features purpose by artificially suppressing it; when they did this and re-ran the prompt, the model instead ended the sentence with the word jaguar. When they kept the rhyming feature but suppressed the word rabbit instead, the model ended the sentence with the features next top choice: habit.Anthropic compares this tool to a microscope for AI. But Olah, who led the research, hopes that one day he can widen the aperture of its lens to encompass not just tiny circuits within an AI model, but the entire scope of its computation. His ultimate goal is to develop a tool that can provide a "holistic account" of the algorithms embedded within these models. I think there's a variety of questions that will increasingly be of societal importance, that this could speak to, if we could succeed, he says. For example: Are these models safe? Can we trust them in certain high-stakes situations? And when are they lying?Universal languageThe Anthropic research also found evidence to support the theory that language models think in a non-linguistic statistical space that is shared between languages.Anthropic scientists tested this by asking Claude for the opposite of small in several different languages. Using their new tool, they analyzed the features that activated inside Claude when it answered each of those prompts in English, French, and Chinese. They found features corresponding to the concepts of smallness, largeness, and oppositeness, which activated no matter what language the question was posed in. Additional features would also activate corresponding to the language of the question, telling the model what language to answer in.This isnt an entirely new findingAI researchers have conjectured for years that language models think in a statistical space outside of language, and earlier interpretability work has borne this out with evidence. But Anthropics paper is the most detailed account yet of exactly how this phenomenon happens inside a model, Olah says.The finding came with a tantalizing prospect for safety research. As models get larger, the team found, they tend to become more capable of abstracting ideas beyond language and into this non-linguistic space. This finding could be useful in a safety context, because a model that is able to form an abstract concept of, say, harmful requests is more likely to be able to refuse them in all contexts, compared to a model that only recognizes specific examples of harmful requests in a single language.This could be good news for speakers of so-called low-resource languages that are not widely represented in the internet data that is used to train AI models. Todays large language models often perform more poorly in those languages than in, say, English. But Anthropics finding raises the prospect that LLMs may one day not need unattainably vast quantities of linguistic data to perform capably and safely in these languages, so long as there is a critical mass big enough to map onto a models internal non-linguistic concepts. However, speakers of those languages will still have to contend with how those very concepts have been shaped by the dominance of languages like English, and the cultures that speak them.Toward a more interpretable futureDespite these advances in AI interpretability, the field is still in its infancy, and significant challenges remain. Anthropic acknowledges that even on short, simple prompts, our method only captures a fraction of the total computation expended by Claudethat is, there is much going on inside its neural network into which they still have zero visibility. It currently takes a few hours of human effort to understand the circuits we see, even on prompts with only tens of words, the company adds. Much more work will be needed to overcome those limitations.But if researchers can achieve that, the rewards might be vast. The discourse around AI today is very polarized, Olah says. At one extreme, there are people who believe AI models "understand" just like people do. On the other, there are people who see them as just fancy autocomplete tools. I think part of whats going on here is, people dont really have productive language for talking about these problems," Olah says. "Fundamentally what they want to ask, I think, is questions of mechanism. How do these models accomplish these behaviors? They dont really have a way to talk about that. But ideally they would be talking about mechanism, and I think that interpretability is giving us the ability to make much more nuanced, specific claims about what exactly is going on inside these models. I hope that that can reduce the polarization on these questions.

The Trump administration is facing heavy blowback for using Signal, a messaging app, to discuss sensitive military plans. On March 24, officials usage of the app was revealed after The Atlantic editor Jeffrey Goldberg published a story titled "The Trump Administration Accidentally Texted Me Its War Plans," in which Secretary of Defense Pete Hegseth, among others, discussed upcoming military strikes on Yemen. The U.S. government previously discouraged federal employees from using the app for official business. Some experts have speculated that sharing sensitive national security details over Signal could be illegal, and Democratic lawmakers have demanded an investigation. If our nation's military secrets are being peddled around over unsecure text chains, we need to know that at once, New York Democrat Chuck Schumer said on the Senate floor.Signal is one of the most secure and private messaging platforms that exists for general public use. But cybersecurity experts argue that the app should not have been used for this level of sensitive communication. Signal is a very robust app: a lot of cybersecurity professionals use it for our communications that we want to protect, says Michael Daniel, president and CEO of the Cyber Threat Alliance and a cybersecurity coordinator under President Obama. But its not as secure as government communications channels. And the use of these kinds of channels increases the risk that something is going to go wrong.Signals StrengthsSignal was launched in 2014, with the goal of creating a privacy-preserving messaging platform in an age of increasing mass surveillance. Signal conversations are protected by end-to-end encryption, a technique that makes it extremely hard for a third party to intercept or decipher private messages. While other messaging tools may collect sensitive personal data, Signal prides itself on securely protecting information such as messaging contacts, frequency, and duration.The app has other privacy features, such as automatically disappearing messages after a set period and preventing screenshots of conversations. Signal data is stored locally on user's devices, not the companys servers. Our goal is that everyone in the world can pick up their device, and without thinking twice about it, or even having an ideological commitment to privacy, use Signal to communicate with anyone they want, Signal President Meredith Whittaker told TIME in 2022.Read More: Signals President Meredith Whittaker Shares Whats Next for the Private Messaging AppOver the last few years, Signal has been used by dissidents and protestors around the world who want to keep their conversations safe from political enemies or law enforcement. In Ukraine, the U.S. Embassy in Kyiv described Signal as critical to their work in its ability to ensure secure, rapid, and easily accessible communications. The app now has 70 million users worldwide, according to the tracking site Business of Apps.Government UseThe usage of Signal for government purposes is more contentious. In 2021, the Pentagon scolded a former official for using Signal, saying that it did not comply with the Freedom of Information Act, which decrees the government has legal obligations to maintain federal records. Goldberg, however, reported this week that the Trump officials Signal chat was set to automatically delete messages after a period of time.Sam Vinograd, who served in former President Barack Obama's Homeland Security Department, told CBS that sharing sensitive security details over Signal could violate the Espionage Act as well. Top intelligence officials testified this week that no classified information was shared over the group chat. CIA Director John Ratcliffe said that Signal was a permissible work-use application for the CIA.Last week, a Pentagon advisory cautioned military personnel against using Signal due to Russian hackers targeting the app.The Cyber Threat Alliances Daniel says that he was surprised that top officials were using Signal, given that they have access to government-specific channels that are more secure. When discussing sensitive information, officials are typically required to do so in designated, secure areas called Sensitive Compartmented Information Facilities (SCIFs), or to use SIPRNet, a secure network used by the Defense and State Departments.These are very senior officials who have a lot of options. They have people whose entire jobs are is to make sure that they're able to communicate at all times, Daniel says. We've had that for decades now, and those procedures are really well honed.Daniel contends that government tools could have prevented what went wrong in this instance: the human error of an outside party mistakenly being added to a message chain. He says that government channels have a much higher level of authentication to ensure that members of communication channels are supposed to have access.Dave Chronister, the CEO of the cybersecurity company Parameter Security, says that the governments bespoke communications channels prevent other kinds of interlopers or hackers attempting to use phishing or malware techniques to learn information. If youre on a cell phone, I dont know who could be looking over my shoulder to see what Im typing, not to mention I dont know what else is on that mobile device, he says.Chronister adds that officials use of Signal, as opposed to internal channels, also makes it harder for the government to identify and contain breaches once theyve happened. We could have data out there we didnt know was compromised, he says. If top cabinet officials are using Signal, Im wondering how much is being done on a daily basisand I think theres going to be a lot more fallout from this.A representative for Signal did not immediately respond to a request for comment.

The genetic testing company 23andMe, which holds the genetic data of 15 million people, declared bankruptcy on Sunday night after years of financial struggles. This means that all of the extremely personal user data could be up for saleand that vast trove of genetic data could draw interest from AI companies looking to train their data sets, experts say. Data is the new oiland this is very high quality oil, says Subodha Kumar, a professor at the Fox School of Business at Temple University. With the development of more and more complicated and rigorous algorithms, this is a gold mine for many companies.But any AI-related company attempting to acquire 23andMe would run significant reputational risks. Many people are horrified by the thought that they surrendered their genetic data to trace their ancestry, only for it to now be potentially used in ways they never consented to.Anybody touching this data is running a risk, Kumar, who is the director of Foxs Center for Business Analytics and Disruptive Technologies, says. But at the same time, not touching it, they might be losing on something big as well.Training LLMsCompanies like OpenAI and Google have poured time and resources into making an impact on the medical field, and 23andMes data trove may attract interest from large AI firms with the financial means to acquire it. 23andMe was valued at around $48 million this week, down from a peak of $6 billion in 2021.These companies are striving to build the most powerful general purpose models possible, which are trained on vast amounts of granular data. But researchers have argued that high-quality data sources are drying up, which makes new and robust information sources all the more coveted. A TechCrunch survey of venture capitalists earlier this year found that more than half of respondents cited the quality or rarity of their proprietary data as the edge that AI startups have over their competition.I think it could be a really valuable data set for some of the big AI companies because it represents this ground truth data of actual genetic data, Kazlauskas says of 23andMe. Some of the human errors that might exist in bio publications, you could avoid.Kumar says that 23andMes data could be especially valuable to companies in their push for agentic AI, or AIs that can perform tasks without the involvement of humans, whether in medical research or company decisionmaking.The whole goal of agentic AI models has been a modular approach: you crack the smaller pieces of the problem and then you put them together, he says.Representatives for Google and OpenAI did not immediately respond to requests for comment.Industry-Based Value23andMes data could also be valuable across different industries using AI to sort through vast amounts of datafirst and foremost, medical research.23andMe already had agreements in place with pharmaceutical companies such as GlaxoSmithKline, which tapped into the companys data sets in the hopes of developing new treatments for disease. Kumar says that at Temple, he and colleagues are working on a project to create personalized treatment for ovarian cancer patientsand have found that genetic data can be very, very powerful in understanding structures that we were not able to understand, he says.However, Alex Zhavoronkov, founder and CEO at Insilico Medicine, contends that 23andMes data may not be as valuable as some think, especially in relation to drug discovery. "Most low hanging fruits have already been picked up and there is significant data in the public domain published together with major academic papers, he wrote in an email to TIME.But companies in many other industries will likely be interested, too. This is an abnormally large and nuanced data set: This amount of genetic data, especially that which comes with personal health and medical records, is rarely publicly accessible, says Anna Kazlauskas, CEO of Open Data Labs and the creator of Vana, a network for user-owned data. All of that contextual data makes it really valuableand hard data to get, she says.Potentially interested industries include insurance companies, who could use the data to identify people with greater health risks, in order to up their premiums. Financial institutions could track the relationship between genetic markers and spending patterns in the process of assessing loans. And e-commerce companies could use the data to tailor ads to people with specific medical conditions.Ethical and Privacy ConcernsBut companies also face significant reputational risks in getting involved. 23andMe suffered a hack in 2023 which exposed the personal data of millions of users, severely hurting the companys reputation. Bidders who come from other industries may have even less data protection than 23andMe did, Kumar says. My worry is that some of the companies are not used to having this kind of data, and they may not have enough governance in place, he says.This is especially dangerous because genetic information is inherently sensitive and cannot be altered once compromised. The genetic information of family members of people who willingly gave their data to the company are also at risk. And given AIs well-known biases, the misuse of such data could lead to discrimination in areas like hiring, insurance and loans. On Friday, California Attorney General Rob Bonta released an urgent alert to 23andMe customers advising them to ask the company to delete their data and destroy their genetic samples under a California privacy law.Eva Galperin, director of cybersecurity at the Electronic Frontier Foundation, worries that 23andMes genetic data might exist in a state of permanent flux on the market. Once you have sold the data, there are no limits to how many times it may be resold, she says. This could result in genetic data falling into the hands of organizations that may not prioritize ethical considerations or have robust data protection measures in place.Insilico Medicines Zhavoronkov says all of these fears mean that potential AI-related bidders will be dissuaded from trying to purchase 23andMe and its data. Their dataset is actually toxic, he says. Whoever buys it and trains on it will get negative publicity, and the acquirer will be possibly investigated or sued."Regardless of what ultimately happens, Kazlauskas says she is at least thankful that this conundrum has opened up larger conversations about data sovereignty. We should probably, in the future, want to avoid this kind of situation where you decide you want to do a genetic test, and then five years later, this company is struggling financially, and that now puts your genetic data at risk of being sold to the highest bidder, she says. In this AI era, that data is super valuable.