The Dos & Don'ts Of Fintech VCgettyAs the founder of a global VC firm, a member of Kauffman Fellows, the global network of venture capitalists, and an investor in a number of startups around the world, Im always astounded that in every deal, different VCs focus on different things. Ive been fortunate to invest in several fintech unicorns from the early stage. Here are ten counterintuitive (and hopefully occasionally controversial) strategies for successful fintech investing.1. The Moat and the TAM Equation For Fintech InvestingConventional wisdom is that moats are key.But financial services is one of the largest markets in the US economy, roughly 20% of GDP. Taking a consumer and small business (SMB) lens for now: most products are not that differentiated over time. Among the largest fintech success stories, many have won because of: a superior product (e.g. simple API with Stripe), better value proposition (free bank account at Chime, a former portfolio company) or unique go to market (e.g. Guideline partnering with Gusto). This can often be complemented by serving an underserved but massive customer segment (e.g. Nubank).My view is that while moat matters and can be defensible over time (particularly for enterprise startups), the Total Addressable Markets (TAMs) in financial services are so large that product differentiation and sales velocity matter even more. Robinhood is a massive business, but with a service that many others offer today. Ramp was not first to market in the spend management category but moved very quickly in launching products.Yes, incumbents (and other startups) will inevitably copy you. But only if youre too slow to grow and iterate. And as we will see in the unit economics section, fintech can generate moats from scale.MORE FOR YOU2. An Engaged Attitude With RegulatorsFintech is Fin first. I personally am wary of any pitch that does not work with a spirit of openness and transparency with the regulators (perhaps it is one reason I have not been a crypto investor over the last decade!) I am of course biased as a former regulator myself.I look for entrepreneurs that do more than comply with the letter of the law, but respect it, and the reason customer protection exists - even if they dont agree with it.The best founders Ive worked with have engaged with regulators with an attitude of openness and engagement.3. Younger Is Not Necessarily Better For Fintech InvestingAs I discussed in Out-Innovate, the stereotype of the hoodie-wearing, 20-something fintech founder is giving way to seasoned entrepreneurs with years of industry experience. Certainly, there are examples of generational founders solving companies young. Stripe for instance is a strong example in fintech (and Meta in social media).However, many successful fintech founders bring domain expertise honed over decades. This is of course, tied to the attitude point above.In fintech, I often overemphasize understanding why a particular founder is uniquely qualified to tackle the particular issue. As a consequence, this biases towards more experienced founders with deep domain expertise.4. Geography Can Be Turned Into An AdvantageNew research by Kauffman Fellows demonstrates the rise of Fund Returners, increasingly coming from outside Silicon Valley and all over the world.Global geographies have a few structural advantages for fintechs. The cost of building a startup is often lower, so the same raise goes meaningfully farther (thus derisking the business).Global startups often have less competition, making success larger. Of course, there are certain drawbacks to building outside Silicon Valley: market size may be smaller, and there is less downstream access to venture capital.Perhaps it is no surprise that the largest neobank in the world is in Brazil (Nubank), the largest input & embedded fintech marketplace is in India (Ofbusiness), and arguably the largest BNPL company globally is in Sweden (Klarna).5. Unit Economics RuleThe shiny allure of rapid growth means little if its built on unsustainable unit economics. Readers of this column will note my preference for camels.When I meet founders, I try to understand what their unit economics are today. Do they have a handle on them? How strong are they? What is the timing of cash flows (see next point)?Clearly, things wont always be perfect, particularly in the early days. Therefore, it is important to understand what pieces the company has a handle on and what still needs to be proven out. How much have the founders researched this? How informed is it with real world experience?Fintech can be a scale game. For example, payment companies and neobanks can consistently get much better deals from partners simply by being bigger. If unit economics are passable in the early stage, there is actually a consistent path to improve. This is different than many other categories where unit economics necessarily degrade over time. However, unit economincs at the beginning are necessarily the foundation.6. Cash Cycles MatterA good fintech will have strong unit economics. A great one will also have a negative cash cycle.To make a simple example, if a lending startup receives an 80% advance rate vs 100% (the percentage of the loan that can be financed), this can be the difference between having to fundraise or notor between life or death. Pardoxically, growth can actually hurt the business. If a lending company sees explosive volume and deploys $100m, in the first scenario, they will need to raise $20m just to be in business (on top of whatever it costs them to set up the operation). In the latter, it is $0 in incremental equity. In short, in the former scenario, as the company grows it will require more cash. In the latter, it will at least be neutral.The same dynamic plays out elsewhere in fintech. An insuretech, may need to fund a portion of their book (especially if a full stack carrier) and a payment company or neobank could need reserve balances.In some cases, a fintech can structure cash cycle deals where they get paid upfront, and only need to pay their counterparties in a few days. With enough growth, this can help fund the business. Nubank for instance has benefited from a negative cash cycle the company typically gets cash upfront from customers and pays its merchants a little later. This can help accelerate growth.Optimizing the cash cycle of a company can make a good fintech an excellent one.7. A Fintech Should Have A Financial Model. Yes, Even At Seed StageAn early-stage financial model is by nature an exercise in futility. Startups, afterall are not companies. They are projects in search of a business model. They do not yet know that business model with certainty at preseed and only start to define it around Series A (and sometimes well beyond). Unit economics, depth of sales channels, etc are at best educated guesses in the early days.As a result, many founders and VCs argue that a financial model is not required by companies at this stage.For fintech I disagree: most startups live and die some form of financial intermediation. A lending startup makes money by first giving it away, losing a portion, and paying others for the use of their capital. Insurance companies get paid premiums, pay out risks later, and manage policies in between.While the ultimate inputs will certainly not be perfect, I evaluate a founders ability to forecast, the thoughtfulness of their approach, and the scalability of the model. Do they have a point of view on key metrics like cost of good sold, gross margin, or cash burn? How nuanced is their perspective based on different scenarios.Every VC tries to understand how strong a founders grasp is on the domain and model. The financial model is a key way to showcase this.8. VCs should not make their own modelsGiven that early-stage founders dont have certainty on the business model or the assumptions that underpin them, the model exercise by nature is imperfect.Compounding imperfection by recreating a new venture capital model is simply compounding errors (this changes at the later stage, of course, when there is more certainty on the model assumptions, and thus more value in creating scenarios). After all, a venture capitalist is necessarily less close to the business, the assumptions, and the operations. How could they do better?As a policy, we do not build our own models for seed stage companies. We leverage the founders model, and then build sensitivities on top of it. We do spend a lot of time understanding the impact of unit economics assumptions, dilution, and exit multiples on the investment return.9. Business Model Depth Not BuzzwordsThe last few years of fintech have been awash by buzzword after buzzword: Crypto. NFT. And now AI.No tech trend will be a panacea. Early enthusiasm is often overblown.Financial services are massive markets. Despite the buzz of fintech, they remain dominated by incumbents. The next big trend may be hot, but this doesnt mean it will conquer fintech.A founder doesnt need to pitch reinventing everything. Partnering or working with incumbents (with data and distribution) is powerful.To be right in venture requires being contrarian (and being right). As a result, a VC should care less about whats hot today, and more about what has a chance to make enduring change within the existing context.10. The customerA focus on a customer is required for every VC. In some ways it is even more important in fintech.If youre wondering whether the regulator will block something, a heuristic I find valuable is asking oneself: is this actually good for the customer? And is the customer being treated like an adult, with full transparency, openness, and dignity?While this isnt surefire, this has generally worked well. Companies that do something fundamentally good for customersprovide a service at a lower cost, or with greater transparencyshould exist.Financial services are quite sticky. If youre wondering if a customer will make the effort to switch, ask yourself if the product is really that much better. Ask yourself if you, your wife, or your mom would use it. You must be able to give a resounding yes to this answer.The customers voice is paramount for us.ConclusionFintech is not like other industries. A different VC playbook matters. Here were a few in mine. What am I missing?

To thrive in this new era, software developers should cultivate a diverse skill set.

A hot potato: Microsoft is raising eyebrows after announcing that it will host DeepSeek R1 on its Azure cloud service. The decision comes just days after OpenAI accused DeepSeek of violating its terms of service by allegedly using ChatGPT outputs to train its system, allegations Microsoft is currently investigating. DeepSeek R1 began making waves in the AI world when it launched last week. Chinese developer DeepSeek touted it as a freely available simulated reasoning model that rivals OpenAI's o1 in performance but at a fraction of the training cost. While OpenAI has priced its o1 model at $60 per million output tokens, DeepSeek lists R1 at just $2.19 per million a remarkable contrast that sunk stock for AI-adjacent companies like Nvidia.Microsoft's decision to host R1 on Azure is not too unusual on its surface. The tech giant already offers over 1,800 AI models through its Azure AI Foundry, giving developers access to a variety of AI systems for experimentation and integration. Microsoft doesn't discriminate since it profits from any AI platform operating on its cloud infrastructure. However, the decision seems ironic since OpenAI has spent the last week aggressively criticizing the model for distilling ChatGPT outputs.OpenAI claims the AI startup violated its terms of service by using "distillation," as reported by Fox News. Distillation is when developers train an AI model using outputs from a more advanced system. Suspicions arose after users discovered that an earlier model, DeepSeek V3, sometimes referred to itself as "ChatGPT," suggesting that DeepSeek used OpenAI-generated data to fine-tune its system.The move also seems somewhat hypocritical, considering Microsoft security researchers reportedly launched an ethics probe into DeepSeek, on Wednesday. Anonymous sources claim that the investigation focuses on whether DeepSeek extracted substantial amounts of data through OpenAI's API during the fall of 2024. // Related StoriesDespite the frustrations with DeepSeek, OpenAI CEO Sam Altman has publicly welcomed the competition. In a tweet on Monday, Altman acknowledged R1's cost efficiency, calling it "an impressive model" but vowing that OpenAI would soon deliver "much better results." Analysts expect the company may release a new model, o3-mini, as early as today.OpenAI's outcry over DeepSeek's data practices is notable given its own history of alleged data abuse. The New York Times has filed a lawsuit against OpenAI and Microsoft, accusing them of using copyrighted journalism without permission. OpenAI has also struck deals with publishers and online communities such as The associated Press and others to access user-generated data for training.The whole situation exposes the AI industry's hypocritical relationship with data ownership. Investment firm Andreessen Horowitz, another Open AI investor, argued in a 2023 legal filing that training AI models should not be considered copyright infringement, as they merely "extract information" from existing works. If OpenAI truly believes in that principle, then DeepSeek is just playing by the same rules.The current landscape of the AI industry is more or less a free-for-all. We have no laws on the books to govern AI directly, and those laws that affect it indirectly, like copyright and trade laws, are twisted into a favorable interpretation by the AI firms that are breaking them.

For one of the best laptop deals around today, look no further than Lenovo. The popular brand offers a great range of business and gaming laptops, with the highlight right now being on its ThinkPad range. Currently, you can buy the Lenovo ThinkPad P16s laptop for a huge 43% off, bringing it down to $2,099 for some great high-end hardware. According to Lenovo, it normally costs $3,689, which explains the $1,590 discount. Granted, Lenovo is known for using an estimated value system sometimes inflates the original price. However, that doesnt change the fact that this is a great price for whats offered here. Lets dip into the finer details before you decide to buy.As mentioned, Lenovo is known for making a lot of business laptops. Crucially, its one of the best laptop brands around in this field thanks to being highly robust and reliable. With this particular model, you get an Intel Core Ultra 7 155H CPU teamed up with 32GB of RAM and 1TB of SSD storage. It also has a dedicated graphics card with the Nvidia RTX 500 Ada GPU, which is roughly comparable to the GeForce RTX 3050 Ti but crucially is designed with creative purposes in mind rather than gaming. If youre looking to render models and videos, this could be useful.Working alongside that GPU is a 16-inch WQUXGA screen with 3840 x 2400 resolution. Its an OLED panel, which means you get the deepest blacks and most vibrant colors all on one screen. This is the technology that many of the best TVs embrace and is the kind of thing youll never want to switch away from. The screen also has Dolby Vision support, HDR 500 True Black, 100% DCI-P3, and 400 nits of brightness. It looks great, with its only minor downfall being a basic 60Hz refresh rate.RelatedAnother reason this might be one of the best laptops for you is the webcam: a 5MP model with RGB+IR, along with a privacy shutter and dual microphones. Theres also a backlit keyboard with a fingerprint reader built-in. The laptop runs Windows 11 Pro our comparison piece between Windows 11 Pro and Windows 11 Home will help you understand why this is the best option for business.Packed with all the key features you need, the Lenovo ThinkPad P16s is usually $3,689 according to Lenovos estimated value system. Its now down to $2,099, which is a great price for a highly competent business laptop. Check out this great business laptop deal for yourself at Lenovo. It could be just what your business needs.Editors Recommendations

Epson Expression Photo XP-8800MSRP$299.99 Score Details The Epson Expression Photo XP-8800 is a surprisingly low-cost printer with a six-color ink system for lab-quality photo prints.ProsGreat low priceExcellent photo quality1200 dpi scannerThree paper traysSD card and USB drive supportIt's a tiny printerConsSlow document printsInk cartridge costs add up quicklyShiny black panels collect fingerprintsTable of ContentsTable of ContentsSpecsDesignPrinting performanceSpecial featuresSoftware and compatibilityPriceIs this the printer for you?Epson is known for high-quality photo printers, so when the company introduced the new Expression Photo XP-8800, I knew I had to test it out. This low-cost all-in-one solution could be ideal for the home, particularly if youre a photography enthusiast.Recommended VideosI checked printer and scanner quality as well as performance across a range of photos and documents. I also researched long-term costs, a critical consideration when shopping for the best photo printer.Epson Expression Photo XP-8800Dimensions13.7 x 20.7 x 7.2 inchesWeight18 poundsPrint speed9.5 ppm (black), 9 ppm (color)Copy speed9 cpm (black), 8 cpm (color)Print resolution5760 x 1440 dpiScan resolution1200 x 4800 dpiPortsHi-Speed USBPaper capacity100 sheets (main), 20 photos (media), single sheet (rear)Wi-FiWi-Fi 802.11b/g/nThe Expression Photo XP-8800 has three paper trays including a convenient media tray. Alan Truly / Digital TrendsThe Expression Photo XP-8800 is tiny for an all-in-one, with a footprint of 13.7 by 13.4 inches and standing just 5.6 inches tall. The output bin is a tray that automatically slides out about seven inches to catch pages and photos when printing. Epson packed a lot of technology into a small package, but the weight is under 15 pounds, so its easy to handle.Speaking of handling, the shiny black control panel and sleek black body look nice but will be a magnet for fingerprints. I kept a microfiber cloth nearby to wipe it clean before taking photos. A large 4.3-inch color touchscreen makes navigation easy.The Expression Photo XP-8800s color display shows large previews and is easy to navigate. Photo by Tracey Truly / Digital TrendsThe Expression Photo XP-8800 can print directly from thumb drives and SD cards via a card slot behind the control panel and a USB-A port behind the paper tray door. The card slot supports SD, SDHC, SDXC, and smaller cards with SD adapters.Epson gave the Expression Photo XP-8800 great versatility. The main paper tray at the bottom holds 100 sheets. Above that, a thin media tray holds 20 photo sheets and doubles as a CD/DVD tray to print directly on rigid discs. The rear tray accepts a variety of paper types, including Epsons Velvet Fine Art Paper and custom paper sizes up to 8.5 x 47.2.As an all-in-one, the Expression Photo XP-8800 includes a flatbed scanner. I can scan documents and photos at up to 1200 dots per inch (dpi). Thats as crisp as some of the best scanners and a good resolution to use when archiving pictures.Epsons Expression Photo XP-8800 excels at photo prints and has great color document quality. Photo by Tracey Truly / Digital TrendsAccording to Epson, the Expression Photo XP-8800 offers Lab-quality photos and prints. In my tests, I found the six-color Claria ink system offered great dynamic range and good color accuracy, confirming that bold statement for a low-cost printer. I tested on several media types from borderless plain paper to glossy 46 photo paper prints. Everything came out looking great.I feel Epsons more expensive EcoTank ET-8500 offered even better picture quality but the biggest difference between the two is performance. The Expression Photo XP-8800 is rated at 9.5 pages per minute (ppm) in black-and-white and 9 ppm in color, about 40% slower than the EcoTank ET-8500.The Expression Photo XP-8800s draft mode is faster and nearly identical to the best quality. Photo by Tracey Truly / Digital TrendsSince the Expression Photo XP-8800 is optimized for photos, thats a more important speed test. A draft mode 46 photo rolls out in about 10 seconds and looks almost identical to high-quality mode. I checked with a macro camera and struggled to see a difference. At full quality, it takes about a minute to print a 4 x 6 photo, so its worth exploring the fastest print option.I loaded a stack of 20 photo cards in the media tray while the main tray held 100 standard letter sheets. A rear cover opens to accept envelopes and specialty paper. With so many options, I didnt have to fuss with paper until it ran low.The Expression Photo XP-8800 supports up to 1200 dpi scanning from a computer. Photo by Tracey Truly / Digital TrendsThe Expression Photo XP-8800 includes a high-quality flatbed scanner that can copy and scan photos or documents. Its about the same speed as the printer, quick enough for home use but could cause a slowdown in an office. It also lacks an automatic document feeder (ADF), so multi-page documents require manual insertion.If you want a business-oriented inkjet printer, take a look at Epsonssuper-quick EcoTank Pro ET-5850. It has very good photo quality, shoots out the first page in 5.5 seconds, and copies double-sided color documents at 20 ppm.The copy resolution is 600 by 600 dpi. Scans offer higher resolution from a computer, I could scan photos at up to 1200 dpi, providing enough size for cropping and fine adjustments. From a phone or tablet, the maximum resolution is 600 dpi, which is still very sharp. Copies look nice and menu options let me reduce and enlarge as needed.The USB port and SD card slots are on the front of the Expression Photo XP-8800 behind the panels. Photo by Tracey Truly / Digital TrendsThe Expression Photo XP-8800 has both a USB-A and an SD card slot, making walk-up printing simple. I could pop a card from my camera directly in the printer and use the touchscreens gallery view to select a print or copy a document to a thumb drive and use that for printing without the need for a phone or computer.Epsons Expression Photo XP-8800 holds six ink cartridges. Alan Truly / Digital TrendsEpson gave the Expression Photo XP-8800 two installation methods: using a mobile app or navigating the touchscreen. I scanned a QR code in the quick start guide to install the Epson Smart Panel app on my iPhone. Its also available for Android phones. The app instantly found the printer and guided me through the setup.The Expression Photo XP-8800 comes with six ink cartridges and I followed Smart Panels instructions to install them. Everything is clearly marked and the cartridges snapped in with reassuring clicks. The app advised me that initialization takes about 10 minutes so I stepped away while it finished up.I ran into a minor issue when the Smart Panel app lost connection to the Expression Photo XP-8800. Photo by Tracey Truly / Digital TrendsWhen I returned, the app had lost connection to the printer, so I used the touchscreen to complete alignment. The Expression Photo XP-8800 printed a single page to place in the scanner for fine adjustments to the printheads to ensure the best quality.I tried the mobile app again to complete the setup and it worked. I connected the Expression Photo XP-8800 to Wi-Fi. My Windows PC, MacBook, and Android phone saw the printer quickly and I had no further issues.From the Epson Smart Panel app, I could print, scan, and copy. Epsons mobile app also supports printing envelopes, something many printers struggle with. It also offers maintenance options to check ink levels, update firmware,While the Epson Expression Photo XP-8800 is very affordable at the $300 retail price, it can be found on sale for as low as $200. Thats an incredible bargain for a six-color all-in-one inkjet printer. The only consideration for this cartridge-based printer is ink cost.It uses Epson Claria 340 ink, available in standard and high-capacity cartridges in black, cyan, magenta, yellow, light cyan, and light magenta. Standard cartridges cost $17 each and provide up to 240 monochrome pages and 360 color pages. High-capacity cartridges cost $28 for black ink and $32 for each color ink. Its refreshing to see the color cartridges have higher yields than the black cartridges. Its usually the other way around.Breaking down those costs, every monochrome page costs 6 to 7 cents and each page of a color document will average 19 to 24 cents. Color costs are much higher than black since multiple inks are used for each dot, while monochrome uses only black.The cost of printing photos is harder to estimate since the content is so varied. Full-page photos could cost significantly more than the average document which often includes plenty of white space. 46 photos are smaller but cover the entire paper with ink.An inkjet tank printer like the Epson EcoTank 2850 costs more upfront but offers great long-term savings with costs measured in tenths of cents per page. Epsons EcoTank ET-8500 is expensive but combines tank printer savings with six-color photo printer quality.The Expression Photo XP-8800 has a compelling price. Its a wonderful compact printer with photographic quality thats sure to please photographers. Just watch the ongoing cost of ink, a somewhat hidden expense of cartridge-based printers.There are plenty of top-quality photo printers that could meet your needs better. If the price is what attracted you to the Expression Photo XP-8800, a great deal on a good printer could handle the job for less.If you want to share a few high-quality pictures with family and friends and have an all-around great printer that can double as a scanner and color copier, Epsons Expression Photo XP-8800 might be the perfect solution.Editors RecommendationsThe best all-in-one printers you can buy in 2024

Waste Management will shed 1,000 jobs this year, accelerating its automation efforts and moving the company closer to its goal of $450 million in cost cuts.

The company has heavily advertised AI features since the latest iPhones were released in September.

Amid a number of thin, depressing movies at this years edition of the Utah festival, a pair of funny, big-hearted films stood out, alongside the typical contingent of thoughtful character studies and powerful documentaries.

Gemma Arterton reprises her winning performance as a British TV star in the latest season of this feisty, funny comedy on PBS.

Control issues Copyright Office suggests AI copyright debate was settled in 1965 Most people think purely AI-generated works shouldn't be copyrighted, report says. Ashley Belanger Jan 30, 2025 3:46 pm | 22 Ars used Copilot to generate this AI image using the precise prompt the Copyright Office used to determine that prompting alone isn't authorship. Credit: AI image generated by Copilot Ars used Copilot to generate this AI image using the precise prompt the Copyright Office used to determine that prompting alone isn't authorship. Credit: AI image generated by Copilot Story textSizeSmallStandardLargeWidth *StandardWideLinksStandardOrange* Subscribers only Learn moreThe US Copyright Office issued AI guidance this week that declared no laws need to be clarified when it comes to protecting authorship rights of humans producing AI-assisted works."Questions of copyrightability and AI can be resolved pursuant to existing law, without the need for legislative change," the Copyright Office said.More than 10,000 commenters weighed in on the guidance, with some hoping to convince the Copyright Office to guarantee more protections for artists as AI technologies advance and the line between human- and AI-created works seems to increasingly blur.But the Copyright Office insisted that the AI copyright debate was settled in 1965 after commercial computer technology started advancing quickly and "difficult questions of authorship" were first raised. That was the first time officials had to ponder how much involvement human creators had in works created using computers.Back then, the Register of Copyrights, Abraham Kaminsteinwho was also instrumental in codifying fair usesuggested that "there is no one-size-fits-all answer" to copyright questions about computer-assisted human authorship. And the Copyright Office agrees that's still the case today."Very few bright-line rules are possible," the Copyright Office said, with one obvious exception. Because of "insufficient human control over the expressive elements" of resulting works, "if content is entirely generated by AI, it cannot be protected by copyright."The office further clarified that doesn't mean that works assisted by AI can never be copyrighted."Where AI merely assists an author in the creative process, its use does not change the copyrightability of the output," the Copyright Office said.Following Kaminstein's advice, officials plan to continue reviewing AI disclosures and weighing, on a case-by-case basis, what parts of each work are AI-authored and which parts are human-authored. Any human-authored expressive element can be copyrighted, the office said, but any aspect of the work deemed to have been generated purely by AI cannot.Prompting alone isnt authorship, Copyright Office saysAfter doing some testing on whether the same exact prompt can generate widely varied outputs, even from the same AI tool, the Copyright Office further concluded that "prompts do not alone provide sufficient control" over outputs to allow creators to copyright purely AI-generated works based on highly intelligent or creative prompting.That decision could change, the Copyright Office said, if AI technologies provide more human control over outputs through prompting.New guidance noted, for example, that some AI tools allow prompts or other inputs "to be substantially retained as part of the output." Consider an artist uploading an original drawing, the Copyright Office suggested, and prompting AI to modify colors, or an author uploading an original piece and using AI to translate it. And "other generative AI systems also offer tools that similarly allow users to exert control over the selection, arrangement, and content of the final output." The Copyright Office drafted this prompt to test artists' control over expressive inputs that are retained in AI outputs. Credit: Copyright Office "Where a human inputs their own copyrightable work and that work is perceptible in the output, they will be the author of at least that portion of the output," the guidelines said.But if officials conclude that even the most iterative prompting doesn't perfectly control the resulting outputseven slowly, repeatedly prompting AI to produce the exact vision in an artist's headsome artists are sure to be disappointed. One artist behind a controversial prize-winning AI-generated artwork has staunchly defended his rigorous AI prompting as authorship.However, if "even expert researchers are limited in their ability to understand or predict the behavior of specific models," the Copyright Office said it struggled to see how artists could. To further prove their point, officials drafted a lengthy, quirky prompt about a cat reading a Sunday newspaper to compare different outputs from the same AI image generator. Copyright Office drafted a quirky, lengthy prompt to test creative control over AI outputs. Credit: Copyright Office Officials apparently agreed with Adobe, which submitted a comment advising the Copyright Office that any output is "based solely on the AIs interpretation of that prompt." Academics further warned that copyrighting outputs based only on prompting could lead copyright law to "effectively vest" authorship adopters with "rights in ideas.""The Office concludes that, given current generally available technology, prompts alone do not provide sufficient human control to make users of an AI system the authors of the output. Prompts essentially function as instructions that convey unprotectable ideas," the guidance said. "While highly detailed prompts could contain the users desired expressive elements, at present they do not control how the AI system processes them in generating the output."Hundreds of AI artworks are copyrighted, officials sayThe Copyright Office repeatedly emphasized that most commenters agreed with the majority of their conclusions. Officials also stressed that hundreds of AI artworks submitted for registration, under existing law, have been approved to copyright the human-authored elements of their works. Rejections are apparently expected to be less common."In most cases," the Copyright Office said, "humans will be involved in the creation process, and the work will be copyrightable to the extent that their contributions qualify as authorship."For stakeholders who have been awaiting this guidance for months, the Copyright Office report may not change the law, but it offers some clarity.For some artists who hoped to push the Copyright Office to adapt laws, the guidelines may disappoint, leaving many questions about a world of possible creative AI uses unanswered. But while a case-by-case approach may leave some artists unsure about which parts of their works are copyrightable, seemingly common cases are being resolved more readily. According to the Copyright Office, after each decision, it gets easier to register AI works that meet similar standards for copyrightability. Perhaps over time, artists will grow more secure in how they use AI and whether it will impact their exclusive rights to distribute works.That's likely cold comfort for the artist advocating for prompting alone to constitute authorship. One AI artist told Ars in October that being denied a copyright has meant suffering being mocked and watching his award-winning work freely used anywhere online without his permission and without payment. But in the end, the Copyright Office was apparently more sympathetic to other commenters who warned that humanity's progress in the arts could be hampered if a flood of easily generated, copyrightable AI works drowned too many humans out of the market."We share the concerns expressed about the impact of AI-generated material on human authors and the value that their creative expression provides to society. If a flood of easily and rapidly AI-generated content drowns out human-authored works in the marketplace, additional legal protection would undermine rather than advance the goals of the copyright system. The availability of vastly more works to choose from could actually make it harder to find inspiring or enlightening content."New guidance likely a big yawn for AI companiesFor AI companies, the copyright guidance may mean very little. According to AI company Hugging Face's comments to the Copyright Office, no changes in the law were needed to ensure the US continued leading in AI innovation, because "very little to no innovation in generative AI is driven by the hope of obtaining copyright protection for model outputs."Hugging Face's Head of ML & Society, Yacine Jernite, told Ars that the Copyright Office seemed to "take a constructive approach" to answering some of artists' biggest questions about AI."We believe AI should support, not replace, artists," Jernite told Ars. "For that to happen, the value of creative work must remain in its human contribution, regardless of the tools used."Although the Copyright Office suggested that this week's report might be the most highly anticipated, Jernite said that Hugging Face is eager to see the next report, which officials said would focus on "the legal implications of training AI models on copyrighted works, including licensing considerations and the allocation of any potential liability.""As a platform that supports broader participation in AI, we see more value in distributing its benefits than in concentrating all control with a few large model providers," Jernite said. "Were looking forward to the next part of the Copyright Offices Report, particularly on training data, licensing, and liability, key questions especially for some types of output, like code."Ashley BelangerSenior Policy ReporterAshley BelangerSenior Policy Reporter Ashley is a senior policy reporter for Ars Technica, dedicated to tracking social impacts of emerging policies and new technologies. She is a Chicago-based journalist with 20 years of experience. 22 Comments