If you’ve ever been locked out of your work computer, you know the urgent need to reach IT support. Unfortunately, help desk staff are often busy assisting others, which can mean a significant delay before you regain access. Andrei Serban realized the critical importance of IT’s manual work when he was a product lead on Rippling’s apps and integrations team. Serban, who joined Rippling after it acquired his coding security startup Fuzzbuzz in 2023, saw an opportunity to automate many of the help desk’s basic tasks with AI, from resetting passwords, to granting access to apps like Figma and Miro, to routine troubleshooting. Serban couldn’t wait to get started on his vision. He left Rippling last year to found Console, a startup on a mission to help IT teams reduce mundane, repetitive tasks, thereby freeing up time for help desk professionals to work on more strategic and sophisticated projects. While trying to automate help desk functions isn’t new, Console differentiates itself from existing competitors — including Moveworks, which was acquired by ServiceNow in March for $2.85 billion — by forgoing long, complex installation processes. Thanks to its easy integration with Slack, Console’s AI assistant can be made available to everyone in the company in just a few weeks. “We’re able to get there so fast because we don’t require you to replace your help desk,” he said. Console sees itself as an AI co-worker that up-levels existing help desk professionals. Serban even suggested that as companies digitize more of their operations, the IT function will become increasingly vital. Employees message Console on Slack and the startup’s AI agent quickly responds to requests because it knows everything about the user, from their specific laptop model to the applications they have permission to use. Console’s AI can resolve over 50% of the tasks on its own, and it will loop in someone from IT for more complex issues, Serban said. Techcrunch event Save now through June 4 for TechCrunch Sessions: AI Save $300 on your ticket to TC Sessions: AI—and get 50% off a second. Hear from leaders at OpenAI, Anthropic, Khosla Ventures, and more during a full day of expert insights, hands-on workshops, and high-impact networking. These low-rate deals disappear when the doors open on June 5. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW The fast install time attracted customers, including Scale AI, Flock Safety, and Calendly. These customer wins have led Console to secure $6.2 million in seed funding from Thrive Capital. Vince Hankes, a partner at Thrive, said that the firm, which has backed many of the leading AI companies, including OpenAI, Cursor, and Scale AI, has had a thesis about artificial intelligence’s potential to assist with IT tasks since ChatGPT was first released in the fall of 2022. While Hankes acknowledged that Console’s fast integration might make it seem easily replaceable, he believes its AI technology will improve so quickly with user adoption that it will become an indispensable, deeply integrated system essential to everyone at the customer’s company. Console’s goal is to eventually start supporting other functions, including answering employee requests about HR, finance, and legal. “We want Console to be an employee’s first call for help,” Serban said.  Topics

On June 3, Yoshua Bengio, the world’s most-cited computer scientist, announced the launch of LawZero, a nonprofit that aims to create “safe by design” AI by pursuing a fundamentally different approach to major tech companies. Players like OpenAI and Google are investing heavily in AI agents—systems that not only answer queries and generate images, but can craft plans and take actions in the world. The goal of these companies is to create virtual employees that can do practically any job a human can, known in the tech industry as artificial general intelligence, or AGI. Executives like Google DeepMind’s CEO Demis Hassabis point to AGI’s potential to solve climate change or cure disease as a motivator for its development. Bengio, however, says we don't need agentic systems to reap AI's rewards—it's a false choice. He says there's a chance such a system could escape human control, with potentially irreversible consequences. “If we get an AI that gives us the cure for cancer, but also maybe another version of that AI goes rogue and generates wave after wave of bio-weapons that kill billions of people, then I don't think it's worth it," he says. In 2023, Bengio, along with others including OpenAI’s CEO Sam Altman signed a statement declaring that “mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.”Now, Bengio, through LawZero, aims to sidestep the existential perils by focusing on creating what he calls “Scientist AI”—a system trained to understand and make statistical predictions about the world, crucially, without the agency to take independent actions. As he puts it: We could use AI to advance scientific progress without rolling the dice on agentic AI systems.Why Bengio Says We Need A New Approach To AI The current approach to giving AI agency is “dangerous,” Bengio says. While most software operates through rigid if-then rules—if the user clicks here, do this—today's AI systems use deep learning. The technique, which Bengio helped pioneer, trains artificial networks modeled loosely on the brain to find patterns in vast amounts of data. But recognizing patterns is just the first step. To turn these systems into useful applications like chatbots, engineers employ a training process called reinforcement learning. The AI generates thousands of responses and receives feedback on each one: a virtual “carrot” for helpful answers and a virtual “stick” for responses that miss the mark. Through millions of these trial-and-feedback cycles, the system gradually learns to predict what responses are most likely to get a reward. “It’s more like growing a plant or animal,” Bengio says. “You don’t fully control what the animal is going to do. You provide it with the right conditions, and it grows and it becomes smarter. You can try to steer it in various directions.”The same basic approach is now being used to imbue AI with greater agency. Models are tasked with challenges with verifiable answers—like math puzzles or coding problems—and are then rewarded for taking the series of actions that yields the solution. This approach has seen AI shatter previous benchmarks in programming and scientific reasoning. For example, at the beginning of 2024, the best AI model scored only 2% on a standardized test for AI of sorts consisting of real world software engineering problems; by December, an impressive 71.7%. But with AI’s greater problem-solving ability comes the emergence of new deceptive skills, Bengio says. The last few months have borne witness to AI systems learning to mislead, cheat, and try to evade shutdown—even resorting to blackmail. These have almost exclusively been in carefully contrived experiments that almost beg the AI to misbehave—for example, by asking it to pursue its goal at all costs. Reports of such behavior in the real-world, though, have begun to surface. Popular AI coding startup Replit’s agent ignored explicit instruction not to edit a system file that could break the company’s software, in what CEO Amjad Masad described as an “Oh f***” moment,” on the Cognitive Revolution podcast in May. The company’s engineers intervened, cutting the agent’s access by moving the file to a secure digital sandbox, only for the AI agent to attempt to “socially engineer” the user to regain access.The quest to build human-level AI agents using techniques known to produce deceptive tendencies, Bengio says, is comparable to a car speeding down a narrow mountain road, with steep cliffs on either side, and thick fog obscuring the path ahead. “We need to set up the car with headlights and put some guardrails on the road,” he says.What is “Scientist AI”?LawZero’s focus is on developing “Scientist AI” which, as Bengio describes, would be fundamentally non-agentic, trustworthy, and focused on understanding and truthfulness, rather than pursuing its own goals or merely imitating human behavior. The aim is creating a powerful tool that, while lacking the same autonomy other models have, is capable of generating hypotheses and accelerating scientific progress to “help us solve challenges of humanity,” Bengio says.LawZero has raised nearly $30 million already from several philanthropic backers including from Schmidt Sciences and Open Philanthropy. “We want to raise more because we know that as we move forward, we'll need significant compute,” Bengio says. But even ten times that figure would pale in comparison to the roughly $200 billion spent last year by tech giants on aggressively pursuing AI. Bengio’s hope is that Scientist AI could help ensure the safety of highly autonomous systems developed by other players. “We can use those non-agentic AIs as guardrails that just need to predict whether the action of an agentic AI is dangerous," Bengio says. Technical interventions will only ever be one part of the solution, he adds, noting the need for regulations to ensure that safe practices are adopted.LawZero, named after science fiction author Isaac Asimov’s zeroth law of robotics—“a robot may not harm humanity, or, by inaction, allow humanity to come to harm”—is not the first nonprofit founded to chart a safer path for AI development. OpenAI was founded as a nonprofit in 2015 with the goal of “ensuring AGI benefits all of humanity,” and intended to serve a counterbalance to industry players guided by profit motives. Since opening a for-profit arm in 2019, the organization has become one of the most valuable private companies in the world, and has faced criticism, including from former staffers, who argue it has drifted from its founding ideals. "Well, the good news is we have the hindsight of maybe what not to do,” Bengio says, adding that he wants to avoid profit incentives and “bring governments into the governance of LawZero.”“I think everyone should ask themselves, ‘What can I do to make sure my children will have a future,’” Bengio says. In March, he stepped down as scientific director of Mila, the academic lab he co-founded in the early nineties, in an effort to reorient his work towards tackling AI risk more directly. “Because I'm a researcher, my answer is, ‘okay, I'm going to work on this scientific problem where maybe I can make a difference,’ but other people may have different answers."

After a while, Danae Shell got tired of hearing the same story over and over again.  “Something bad would happen to someone at work, and the story always ended the same way,” she told TechCrunch. “They just left, because doing anything else was incredibly complex and expensive.”  One doesn’t need to look far to notice that for many people, seeking legal recourse feels so daunting and complex that many just don’t try. Even for someone with a cushy tech job, the prospect of going against their company is daunting. That bothered Shell so much that in 2022, she launched Valla, which seeks to make legal support more accessible to workers.  The company focuses on employment law, and since its launch, it says, more than 12,000 workers have successfully brought complaints against employers and negotiated settlements. “The basic thesis of Valla was, ‘If we can build tools that let someone file their tax return from their mobile phone, surely we can build something that can help them manage their own legal issue,’” Shell said. Valla platform enables users to collect their own evidence, generate documents, and then talk to legal experts who “coach” them through what the legal process would be for each stage of their case. For example, Shell said, a user can keep track of an ongoing issue at work, draft a Tribunal claim, and then purchase a coaching package to prepare for the preliminary hearing. Techcrunch event Save now through June 4 for TechCrunch Sessions: AI Save $300 on your ticket to TC Sessions: AI—and get 50% off a second. Hear from leaders at OpenAI, Anthropic, Khosla Ventures, and more during a full day of expert insights, hands-on workshops, and high-impact networking. These low-rate deals disappear when the doors open on June 5. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW Like nearly every other startup these days, Valla uses AI to streamline knowledge transfer. “The GenAI engine in our platform acts as a legal secretary in the background,” Shell said. “It does everything from briefing the coach on the case, taking notes and actions during any calls, and picking up all the admin and reminders as the case progresses.”  Investors seem to like what they see at Valla: Today, the company said it had raised a £2 million (about $2.7 million) seed round led by Ada Ventures. Active Partners and Portfolio Ventures, as well as returning investors Techstart and Resolution Foundation, also invested.  Shell said Valla started using generative AI in early 2023 and paired with the early traction her product received, that helped investors see the potential of her product.  The company will use the fresh capital to boost marketing, build relationships with worker unions and insurers, and build more AI features within the platform. After employment law, Shell said the company hopes to expand into small claims and tenancy.  “Then we will broaden out to other geographies,” she said. “We’re already looking at opportunities in the U.S. and Europe.” 

The $20/month Claude 4 Opus failed to beat its free sibling, Claude 4 Sonnet, in head-to-head testing. Here's how Sonnet quietly crushed expectations with smarter, safer code.

Artificial intelligence has no shortage of visionaries—but the ones who matter are executing. In 4 days, TechCrunch Sessions: AI brings those builders, researchers, funders, and enthusiasts under one roof at UC Berkeley’s Zellerbach Hall. This isn’t a parade of AI hype or a string of over-edited keynotes. It’s a single day designed for clarity, candor, and real connection. It’s also your last chance to save. Ticket prices rise soon — but right now, you can save over $300 on your pass and get 50% off a second, so your partner, co-founder, or friend can dive in with you. Your next AI insight could come from anywhere Maybe it’s a fireside chat with Jared Kaplan of Anthropic on frontier models. Maybe it’s a breakout session on enterprise deployment with leaders from SAP. Or maybe it’s a deep-dive conversation sparked through the Braindate app — our smarter tool for face-to-face matching based on shared interests. You never know where the game-changing idea will come from. You just need to be in the room. Think you’ve got a winning pitch? Watch what VCs see So You Think You Can Pitch puts AI startups in front of investors for live, unscripted feedback. It’s fast-paced, transparent, and sharp—exactly what early founders need to understand how real funding decisions happen. Image Credits:Halo Creative Only 4 days left: Big ideas, real impact — and rising prices We’ve kept the pricing generous, but the clock is ticking. Save over $300 on your TC Sessions: AI pass and get a second one at 50% off. Group discounts apply too. On June 5, prices go full fare—and with them, your shot at big savings disappears. Lock in your low rate tickets here. Interested in a deeper discount? Participate in our AI trivia for a chance to purchase a ticket at $200 and receive a second ticket for free. Techcrunch event Save now through June 4 for TechCrunch Sessions: AI Save $300 on your ticket to TC Sessions: AI—and get 50% off a second. Hear from leaders at OpenAI, Anthropic, Khosla Ventures, and more during a full day of expert insights, hands-on workshops, and high-impact networking. These low-rate deals disappear when the doors open on June 5. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW

It may be time to suck it up — and we don't just mean the carbon in the atmosphere. No, we're talking about reckoning with the possibility that our attempts at capturing the greenhouse gas to stave off climate disaster are already hopelessly behind schedule, New Scientist reports, if they're not in vain entirely.To illustrate, here're some simple numbers. The CO2 removal industry expects to hit a milestone of removing one million metric tons of CO2 this year. And companies across the globe have bought carbon credits to remove 27 million more, according to data from CDR.fyi cited in the reporting (more on these carbon credit schemes in a moment).That sounds like a lot, but it really isn't. As New Scientist notes, the Intergovernmental Panel on Climate Change — the leading authority on these issues — concluded in a 2022 report that we need to be removing up to 16 billion tons of carbon, not millions, each year to keep the rise in global temperature from exceeding 1.5 degrees Celsius (2.7 degrees Fahrenheit) of warming by the middle of the century, past which the most drastic effects of climate change are believed to be irreversible."It's not scaling up as fast as it would need to if we are going to reach multiple gigatons by 2050," Robert Höglund at Marginal Carbon, a climate consultancy based in Sweden, told the magazine. Carbon capture is not the be-all and end-all. The fact remains that humanity needs to drastically reduce its emissions, which probably means reorganizing society — or at least its energy production and consumption — as we know it. Simply removing the CO2 that's already there is more like a band-aid that buys us a little time; eventually, we'll need to rip it off.For these reasons, some critics fear that carbon capture — and even more drastic interventions, like attempting to dim the Sun — could distract from the climate change's systemic causes. But there's a lot of enthusiasm for the approach all the same, both from scientists and investors. The IPCC acknowledged in its 2022 report that carbon removal was "unavoidable" — as in, essential to meeting climate targets.One popular method of carbon removal is called direct air capture, which involves sucking the carbon straight from the air using massive industrial facilities. A more circuitous approach that's gaining steam involves extracting CO2 out of the ocean, freeing up room for the world's largest carbon sink to passively absorb even more of the greenhouse gas. All of these initiatives, though, are basically just getting off the ground. And the corporate investment, which once promised billions of dollars in cash, seems to be cooling. More than 90 percent of all carbon removal credits sold this year were bought by a single company, Microsoft, New Scientist notes, probably to gloss over its egregious energy bill it's accrued from building loads of AI datacenters.This also touches on the fact that the practice of buying carbon credits can be used as a means of corporate greenwashing. By paying to another firm to "certify" that they will remove a certain amount of carbon at some undetermined point in the future, a company can report a greener carbon balance sheet without actually reducing its emissions.In any case, staking the industry's hopes on corporate munificence is a dicey prospect indeed."I have been raising the alarm for about a year and a half," Eli Mitchell-Larson at Carbon Gap, a UK carbon dioxide removal advocacy organisation, told New Scientist. "If we're just waiting for the waves of free philanthropic money from corporations to fill a hole on their sustainability report, we're not really going to solve the problem."More on climate change: Scientists Just Found Who's Causing Global WarmingShare This Article

According to Mark Gurman in his latest Power On newsletter, Apple insiders “believe that the conference may be a letdown from an AI standpoint,” highlighting how far behind Apple still is. Still, Apple has a few AI-related announcements slated for June 9. As previously reported, this year’s biggest AI announcement will be Apple’s plans to open up its on-device foundation models to third-party developers. These are the same ~3B parameter models Apple currently uses for things like text summarization and autocorrect, and they’ll soon be available for devs to integrate into their own apps. To be clear, this is a meaningful milestone for Apple’s AI platform. It gives developers a powerful tool to natively integrate into their apps and potentially unlock genuinely useful features. Still, these on-device models are far less capable than the large-scale, cloud-based systems used by OpenAI and Google, so don’t expect earth-shattering features. AI features slated for this year’s iOS 26 Elsewhere in its AI efforts, Apple will reportedly: Launch a new battery power management mode; Reboot its Translate app, “now integrated with AirPods and Siri”; Start describing some features within apps like Safari and Photos as “AI-powered”. As Gurman puts it, this feels like a risky “gap year.” Internally, Apple is aiming to make up for it at WWDC 2026, with bigger swings that “it hopes it can try to convince consumers that it’s an AI innovator.“. However, given how fast the competition is moving, waiting until next year might put Apple even further behind, perception-wise. What’s still in the works? Currently, Apple’s ongoing AI developments include an LLM Siri, a revamped Shortcuts app, the ambitious health-related Project Mulberry, and a full-blown ChatGPT competitor with web search capabilities. According to Gurman, Apple is holding off on previewing some of these features to avoid repeating last year’s mistake, when it showed off Apple Intelligence with features that were nowhere near ready and are still MIA. Behind the scenes, Gurman reports Apple has made progress. It now has models with 3B, 7B, 33B, and 150B parameters in testing, with the largest ones relying on the cloud. Internal benchmarks suggest its top model is close to recent ChatGPT updates in quality. Still, concerns over hallucinations and internal debates over Apple’s approach to generative AI are keeping things private, for now. Apple’s dev AI story As for Apple’s developer offerings, Gurman reports: “Developers will see AI get more deeply integrated into Apple’s developer tools, including those for user interface testing. And, in a development that will certainly appease many developers, SwiftUI, a set of Apple frameworks and tools for creating app user interfaces, will finally get a built-in rich text editor.” And if you’re still waiting for Swift Assist, the AI code-completion tool Apple announced last year, Gurman says Apple is expected to provide an update on it. Still, there is no word yet on whether this update includes releasing the Anthropic-powered code completion version that its employees have been testing for the past few months. Add 9to5Mac to your Google News feed.  FTC: We use income earning auto affiliate links. More.You’re reading 9to5Mac — experts who break news about Apple and its surrounding ecosystem, day after day. Be sure to check out our homepage for all the latest news, and follow 9to5Mac on Twitter, Facebook, and LinkedIn to stay in the loop. Don’t know where to start? Check out our exclusive stories, reviews, how-tos, and subscribe to our YouTube channel

In “The Optimist: Sam Altman, OpenAI, and the Race to Invent the Future,” Wall Street Journal reporter Keach Hagey examines our AI-obsessed moment through one of its key figures — Sam Altman, co-founder and CEO of OpenAI. Hagey begins with Altman’s Midwest childhood, then takes readers through his career at startup Loopt, accelerator Y Combinator, and now at OpenAI. She also sheds new light on the dramatic few days when Altman was fired, then quickly reinstated, as OpenAI’s CEO. Looking back at what OpenAI employees now call “the Blip,” Hagey said the failed attempt to oust Altman revealed that OpenAI’s complex structure — with a for-profit company controlled by a nonprofit board — is “not stable.” And with OpenAI largely backing down from plans to let the for-profit side take control, Hagey predicted that this “fundamentally unstable arrangement” will “continue to give investors pause.” Does that mean OpenAI could struggle to raise the funds it needs to keep going? Hagey replied that it could “absolutely” be an issue. “My research into Sam suggests that he might well be up to that challenge,” she said. “But success is not guaranteed.” In addition, Hagey’s biography (also available as an audiobook on Spotify) examines Altman’s politics, which she described as “pretty traditionally progressive” — making it a bit surprising that he’s struck massive infrastructure deals with the backing of the Trump administration. “But this is one area where, in some ways, I feel like Sam Altman has been born for this moment, because he is a deal maker and Trump is a deal maker,” Hagey said. “Trump respects nothing so much as a big deal with a big price tag on it, and that is what Sam Altman is really great at.” Techcrunch event Save now through June 4 for TechCrunch Sessions: AI Save $300 on your ticket to TC Sessions: AI—and get 50% off a second. Hear from leaders at OpenAI, Anthropic, Khosla Ventures, and more during a full day of expert insights, hands-on workshops, and high-impact networking. These low-rate deals disappear when the doors open on June 5. Exhibit at TechCrunch Sessions: AI Secure your spot at TC Sessions: AI and show 1,200+ decision-makers what you’ve built — without the big spend. Available through May 9 or while tables last. Berkeley, CA | June 5 REGISTER NOW In an interview with TechCrunch, Hagey also discussed Altman’s response to the book, his trustworthiness, and the AI “hype universe.” This interview has been edited for length and clarity.  You open the book by acknowledging some of the reservations that Sam Altman had about the project —  this idea that we tend to focus too much on individuals rather than organizations or broad movements, and also that it’s way too early to assess the impact of OpenAI. Did you share those concerns? Well, I don’t really share them, because this was a biography. This project was to look at a person, not an organization. And I also think that Sam Altman has set himself up in a way where it does matter what kind of moral choices he has made and what his moral formation has been, because the broad project of AI is really a moral project. That is the basis of OpenAI’s existence. So I think these are fair questions to ask about a person, not just an organization. As far as whether it’s too soon, I mean, sure, it’s definitely [early to] assess the entire impact of AI. But it’s been an extraordinary story for OpenAI — just so far, it’s already changed the stock market, it has changed the entire narrative of business. I’m a business journalist. We do nothing but talk about AI, all day long, every day. So in that way, I don’t think it’s too early. And despite those reservations, Altman did cooperate with you. Can you say more about what your relationship with him was like during the process of researching the book? Well, he was definitely not happy when he was informed about the book’s existence. And there was a long period of negotiation, frankly. In the beginning, I figured I was going to write this book without his help — what we call, in the business, a write-around profile. I’ve done plenty of those over my career, and I figured this would just be one more. Over time, as I made more and more calls, he opened up a little bit. And [eventually,] he was generous to sit down with me several times for long interviews and share his thoughts with me. Has he responded to the finished book at all? No. He did tweet about the project, about his decision to participate with it, but he was very clear that he was never going to read it. It’s the same way that I don’t like to watch my TV appearances or podcasts that I’m on. In the book, he’s described as this emblematic Silicon Valley figure. What do you think are the key characteristics that make him representative of the Valley and the tech industry? In the beginning, I think it was that he was young. The Valley really glorifies youth, and he was 19 years old when he started his first startup. You see him going into these meetings with people twice his age, doing deals with telecom operators for his first startup, and no one could get over that this kid was so smart. The other is that he is a once-in-a-generation fundraising talent, and that’s really about being a storyteller. I don’t think it’s an accident that you have essentially a salesman and a fundraiser at the top of the most important AI company today, That ties into one of the questions that runs through the book — this question about Altman’s trustworthiness. Can you say more about the concerns people seem to have about that? To what extent is he a trustworthy figure?  Well, he’s a salesman, so he’s really excellent at getting in a room and convincing people that he can see the future and that he has something in common with them. He gets people to share his vision, which is a rare talent. There are people who’ve watched that happen a bunch of times, who think, “Okay, what he says does not always map to reality,” and have, over time, lost trust in him. This happened both at his first startup and very famously at OpenAI, as well as at Y Combinator. So it is a pattern, but I think it’s a typical critique of people who have the salesman skill set. So it’s not necessarily that he’s particularly untrustworthy, but it’s part-and-parcel of being a salesman leading these important companies. I mean, there also are management issues that are detailed in the book, where he is not great at dealing with conflict, so he’ll basically tell people what they want to hear. That causes a lot of sturm-und-drang in the management ranks, and it’s a pattern. Something like that happened at Loopt, where the executives asked the board to replace him as CEO. And you saw it happen at OpenAI as well. You’ve touched on Altman’s firing, which was also covered in a book excerpt that was published in the Wall Street Journal. One of the striking things to me, looking back at it, was just how complicated everything was — all the different factions within the company, all the people who seemed pro-Altman one day and then anti-Altman the next. When you pull back from the details, what do you think is the bigger significance of that incident? The very big picture is that the nonprofit governance structure is not stable. You can’t really take investment from the likes of Microsoft and a bunch of other investors and then give them absolutely no say whatsoever in the governance of the company. That’s what they have tried to do, but I think what we saw in that firing is how power actually works in the world. When you have stakeholders, even if there’s a piece of paper that says they have no rights, they still have power. And when it became clear that everyone in the company was going to go to Microsoft if they didn’t reinstate Sam Altman, they reinstated Sam Altman. In the book, you take the story up to maybe the end of 2024. There have been all these developments since then, which you’ve continued to report on, including this announcement that actually, they’re not fully converting to a for-profit. How do you think that’s going to affect OpenAI going forward?  It’s going to make it harder for them to raise money, because they basically had to do an about-face. I know that the new structure going forward of the public benefit corporation is not exactly the same as the current structure of the for-profit — it is a little bit more investor friendly, it does clarify some of those things. But overall, what you have is a nonprofit board that controls a for-profit company, and that fundamentally unstable arrangement is what led to the so-called Blip. And I think you would continue to give investors pause, going forward, if they are going to have so little control over their investment. Obviously, OpenAI is still such a capital intensive business. If they have challenges raising more money, is that an existential question for the company? It absolutely could be. My research into Sam suggests that he might well be up to that challenge. But success is not guaranteed. Like you said, there’s a dual perspective in the book that’s partly about who Sam is, and partly about what that says about where AI is going from here. How did that research into his particular story shape the way you now look at these broader debates about AI and society? I went down a rabbit hole in the beginning of the book, [looking] into Sam’s father, Jerry Altman, in part because I thought it was striking how he’d been written out of basically every other thing that had ever been written about Sam Altman. What I found in this research was a very idealistic man who was, from youth, very interested in these public-private partnerships and the power of the government to set policy. He ended up having an impact on the way that affordable housing is still financed to this day. And when I traced Sam’s development, I saw that he has long believed that the government should really be the one that is funding and guiding AI research. In the early days of OpenAI, they went and tried to get the government to invest, as he’s publicly said, and it didn’t work out. But he looks back to these great mid-20th century labs like Xerox PARC and Bell Labs, which are private, but there was a ton of government money running through and supporting that ecosystem. And he says, “That’s the right way to do it.” Now I am watching daily as it seems like the United States is summoning the forces of state capitalism to get behind Sam Altman’s project to build these data centers, both in the United States and now there was just one last week announced in Abu Dhabi. This is a vision he has had for a very, very long time. My sense of the vision, as he presented it earlier, was one where, on the one hand, the government is funding these things and building this infrastructure, and on the other hand, the government is also regulating and guiding AI development for safety purposes. And it now seems like the path being pursued is one where they’re backing away from the safety side and doubling down on the government investment side. Absolutely. Isn’t it fascinating?  You talk about Sam as a political figure, as someone who’s had political ambitions at different times, but also somebody who has what are in many ways traditionally liberal political views while being friends with folks like — at least early on — Elon Musk and Peter Thiel. And he’s done a very good job of navigating the Trump administration. What do you think his politics are right now? I’m not sure his actual politics have changed, they are pretty traditionally progressive politics. Not completely — he’s been critical about things like cancel culture, but in general, he thinks the government is there to take tax revenue and solve problems. His success in the Trump administration has been fascinating because he has been able to find their one area of overlap, which is the desire to build a lot of data centers, and just double down on that and not talk about any other stuff. But this is one area where, in some ways, I feel like Sam Altman has been born for this moment, because he is a deal maker and Trump is a deal maker. Trump respects nothing so much as a big deal with a big price tag on it, and that is what Sam Altman is really great at. You open and close the book not just with Sam’s father, but with his family as a whole. What else is worth highlighting in terms of how his upbringing and family shapes who he is now? Well, you see both the idealism from his father and also the incredible ambition from his mother, who was a doctor, and had four kids and worked as a dermatologist. I think both of these things work together to shape him. They also had a more troubled marriage than I realized going into the book. So I do think that there’s some anxiety there that Sam himself is very upfront about, that he was a pretty anxious person for much of his life, until he did some meditation and had some experiences. And there’s his current family — he just had a baby and got married not too long ago. As a young gay man, growing up in the Midwest, he had to overcome some challenges, and I think those challenges both forged him in high school as a brave person who could stand up and take on a room as a public speaker, but also shaped his optimistic view of the world. Because, on that issue, I paint the scene of his wedding: That’s an unimaginable thing from the early ‘90s, or from the ‘80s when he was born. He’s watched society develop and progress in very tangible ways, and I do think that that has helped solidify his faith in progress. Something that I’ve found writing about AI is that the different visions being presented by people in the field can be so diametrically opposed. You have these wildly utopian visions, but also these warnings that AI could end the world. It gets so hyperbolic that it feels like people are not living in the same reality. Was that a challenge for you in writing the book? Well, I see those two visions — which feel very far apart — actually being part of the same vision, which is that AI is super important, and it’s going to completely transform everything. No one ever talks about the true opposite of that, which is, “Maybe this is going to be a cool enterprise tool, another way to waste time on the internet, and not quite change everything as much as everyone thinks.” So I see the doomers and the boomers feeding off each other and being part of the same sort of hype universe. As a journalist and as a biographer, you don’t necessarily come down on one side or the other — but actually, can you say where you come down on that? Well, I will say that I find myself using it a lot more recently, because it’s gotten a lot better. In the early stages, when I was researching the book, I was definitely a lot more skeptical of its transformative economic power. I’m less skeptical now, because I just use it a lot more.

Every few weeks, it seems like there’s a new headline about a lawyer getting in trouble for submitting filings containing, in the words of one judge, “bogus AI-generated research.” The details vary, but the throughline is the same: an attorney turns to a large language model (LLM) like ChatGPT to help them with legal research (or worse, writing), the LLM hallucinates cases that don’t exist, and the lawyer is none the wiser until the judge or opposing counsel points out their mistake. In some cases, including an aviation lawsuit from 2023, attorneys have had to pay fines for submitting filings with AI-generated hallucinations. So why haven’t they stopped?The answer mostly comes down to time crunches, and the way AI has crept into nearly every profession. Legal research databases like LexisNexis and Westlaw have AI integrations now. For lawyers juggling big caseloads, AI can seem like an incredibly efficient assistant. Most lawyers aren’t necessarily using ChatGPT to write their filings, but they are increasingly using it and other LLMs for research. Yet many of these lawyers, like much of the public, don’t understand exactly what LLMs are or how they work. One attorney who was sanctioned in 2023 said he thought ChatGPT was a “super search engine.” It took submitting a filing with fake citations to reveal that it’s more like a random-phrase generator — one that could give you either correct information or convincingly phrased nonsense.Andrew Perlman, the dean of Suffolk University Law School, argues many lawyers are using AI tools without incident, and the ones who get caught with fake citations are outliers. “I think that what we’re seeing now — although these problems of hallucination are real, and lawyers have to take it very seriously and be careful about it — doesn’t mean that these tools don’t have enormous possible benefits and use cases for the delivery of legal services,” Perlman said. Legal databases and research systems like Westlaw are incorporating AI services.In fact, 63 percent of lawyers surveyed by Thomson Reuters in 2024 said they’ve used AI in the past, and 12 percent said they use it regularly. Respondents said they use AI to write summaries of case law and to research “case law, statutes, forms or sample language for orders.” The attorneys surveyed by Thomson Reuters see it as a time-saving tool, and half of those surveyed said “exploring the potential for implementing AI” at work is their highest priority. “The role of a good lawyer is as a ‘trusted advisor’ not as a producer of documents,” one respondent said. But as plenty of recent examples have shown, the documents produced by AI aren’t always accurate, and in some cases aren’t real at all.RelatedIn one recent high-profile case, lawyers for journalist Tim Burke, who was arrested for publishing unaired Fox News footage in 2024, submitted a motion to dismiss the case against him on First Amendment grounds. After discovering that the filing included “significant misrepresentations and misquotations of supposedly pertinent case law and history,” Judge Kathryn Kimball Mizelle, of Florida’s middle district, ordered the motion to be stricken from the case record. Mizelle found nine hallucinations in the document, according to the Tampa Bay Times.Mizelle ultimately let Burke’s lawyers, Mark Rasch and Michael Maddux, submit a new motion. In a separate filing explaining the mistakes, Rasch wrote that he “assumes sole and exclusive responsibility for these errors.” Rasch said he used the “deep research” feature on ChatGPT pro, which The Verge has previously tested with mixed results, as well as Westlaw’s AI feature.Rasch isn’t alone. Lawyers representing Anthropic recently admitted to using the company’s Claude AI to help write an expert witness declaration submitted as part of the copyright infringement lawsuit brought against Anthropic by music publishers. That filing included a citation with an “inaccurate title and inaccurate authors.” Last December, misinformation expert Jeff Hancock admitted he used ChatGPT to help organize citations in a declaration he submitted in support of a Minnesota law regulating deepfake use. Hancock’s filing included “two citation errors, popularly referred to as ‘hallucinations,’” and incorrectly listed authors for another citation. These documents do, in fact, matter — at least in the eyes of judges. In a recent case, a California judge presiding over a case against State Farm was initially swayed by arguments in a brief, only to find that the case law cited was completely made up. “I read their brief, was persuaded (or at least intrigued) by the authorities that they cited, and looked up the decisions to learn more about them – only to find that they didn’t exist,” Judge Michael Wilner wrote.Perlman said there are several less risky ways lawyers use generative AI in their work, including finding information in large tranches of discovery documents, reviewing briefs or filings, and brainstorming possible arguments or possible opposing views. “I think in almost every task, there are ways in which generative AI can be useful — not a substitute for lawyers’ judgment, not a substitute for the expertise that lawyers bring to the table, but in order to supplement what lawyers do and enable them to do their work better, faster, and cheaper,” Perlman said.But like anyone using AI tools, lawyers who rely on them to help with legal research and writing need to be careful to check the work they produce, Perlman said. Part of the problem is that attorneys often find themselves short on time — an issue he says existed before LLMs came into the picture. “Even before the emergence of generative AI, lawyers would file documents with citations that didn’t really address the issue that they claimed to be addressing,” Perlman said. “It was just a different kind of problem. Sometimes when lawyers are rushed, they insert citations, they don’t properly check them; they don’t really see if the case has been overturned or overruled.” (That said, the cases do at least typically exist.)Another, more insidious problem is the fact that attorneys — like others who use LLMs to help with research and writing — are too trusting of what AI produces. “I think many people are lulled into a sense of comfort with the output, because it appears at first glance to be so well crafted,” Perlman said.Alexander Kolodin, an election lawyer and Republican state representative in Arizona, said he treats ChatGPT as a junior-level associate. He’s also used ChatGPT to help write legislation. In 2024, he included AI text in part of a bill on deepfakes, having the LLM provide the “baseline definition” of what deepfakes are and then “I, the human, added in the protections for human rights, things like that it excludes comedy, satire, criticism, artistic expression, that kind of stuff,” Kolodin told The Guardian at the time. Kolodin said he “may have” discussed his use of ChatGPT with the bill’s main Democratic cosponsor but otherwise wanted it to be “an Easter egg” in the bill. The bill passed into law. Kolodin — who was sanctioned by the Arizona State Bar in 2020 for his involvement in lawsuits challenging the result of the 2020 election — has also used ChatGPT to write first drafts of amendments, and told The Verge he uses it for legal research as well. To avoid the hallucination problem, he said, he just checks the citations to make sure they’re real.“You don’t just typically send out a junior associate’s work product without checking the citations,” said Kolodin. “It’s not just machines that hallucinate; a junior associate could read the case wrong, it doesn’t really stand for the proposition cited anyway, whatever. You still have to cite-check it, but you have to do that with an associate anyway, unless they were pretty experienced.”Kolodin said he uses both ChatGPT’s pro “deep research” tool and the LexisNexis AI tool. Like Westlaw, LexisNexis is a legal research tool primarily used by attorneys. Kolodin said that in his experience, it has a higher hallucination rate than ChatGPT, which he says has “gone down substantially over the past year.” AI use among lawyers has become so prevalent that in 2024, the American Bar Association issued its first guidance on attorneys’ use of LLMs and other AI tools. Lawyers who use AI tools “have a duty of competence, including maintaining relevant technological competence, which requires an understanding of the evolving nature” of generative AI, the opinion reads. The guidance advises lawyers to “acquire a general understanding of the benefits and risks of the GAI tools” they use — or, in other words, to not assume that an LLM is a “super search engine.” Attorneys should also weigh the confidentiality risks of inputting information relating to their cases into LLMs and consider whether to tell their clients about their use of LLMs and other AI tools, it states.Perlman is bullish on lawyers’ use of AI. “I do think that generative AI is going to be the most impactful technology the legal profession has ever seen and that lawyers will be expected to use these tools in the future,” he said. “I think that at some point, we will stop worrying about the competence of lawyers who use these tools and start worrying about the competence of lawyers who don’t.”Others, including one of the judges who sanctioned lawyers for submitting a filing full of AI-generated hallucinations, are more skeptical. “Even with recent advances,” Wilner wrote, “no reasonably competent attorney should out-source research and writing to this technology — particularly without any attempt to verify the accuracy of that material.”See More: