The AI & Legal Update #16 - ⭐️ DeepSeek, American AI, Stargate, and the Right Side of History ↵

Hi, friends. Big month. A new AI company rocked the markets with a model almost as good as OpenAI. We have our first copyright and AI model fair use ruling. Regulators keep regulating, and VP Vance packs an AI policy punch. Phew. Read on.

📘 Read time: 8 minutes 7 seconds

AI & Regulatory

🫠 The Consumer Financial Protection Burueau said in January it would will ramp up oversight of AI use in financial institutions. But … will CFPB survive the Trump administration? Its chief told staff this week “not to perform any work tasks.” See GC AI summary of the administration’s position on the agency. Related, a purge of privacy officials under Trump casts uncertainty on data transfer frameworks.

🎯 Speaking of, President Trump’s January Executive Order to maintain U.S. competitiveness in AI development calls for reduced regulatory barriers, while addressing growing concerns about privacy protection and ethical oversight. By late summer, designated advisors and federal agencies must develop an action plan to “sustain and enhance America’s global AI dominance in order to promote human flourishing, economic competitiveness, and national security.”

🤠 The Texas House introduced Texas House Bill 1709 (HB 1709), the Texas Responsible Artificial Intelligence Governance Act. It would create one of the most comprehensive AI regulatory frameworks in the US. See GC AI summary below.

🗳️ The New Jersey Attorney General’s Office, Division on Civil Rights joins other AGs in releasing guidance on algorithmic discrimination in AI. The guidance clarifies that the New Jersey Law Against Discrimination applies to employers using AI.

Big Tech

🌐 Google released its responsible AI progress report, detailing the company’s “methods for governing, mapping, measuring, and managing AI risks aligned to the NIST framework” and operational updates. Your editor has seen Gemini in her email and is using it - overall, mixed reviews. After all, your editor writes artisan emails that only the very best AI (ahem, GC AI, ahem) can help with. 💁🏻‍♀️

🚶 Microsoft can chew gum and walk at the same time. The company explains: “we are ready to help our customers do two things at once: innovate with AI and comply with the EU AI Act.” See their blog with insights into the company’s prep for enhanced regulatory oversight, including around algorithmic transparency and high-risk AI systems.

🖼️ Open AI has rebranded with a new look and feel, emphasizing the emotional side of AI, with elements of nature and human connection.

🤖 OpenAI came out with a new model Operator, which introduces web browsing and task execution capabilities to take action on your behalf. It takes screenshots every few seconds that the AI analyzes visually. Think: low frame rate TV. Editor's note: We benchmarked Operator against GC AI on some agentic legal searches. For this one, reviewing minimum wage laws in different locations, GC AI answered in one minute. Operator took … an hour and a half and got it wrong! GC AI for the win.

💰 OpenAI, SoftBank, and Oracle are in a $500B partnership for data center development. Known as Stargate, the project intensifies AI model training competition while raising questions about market concentration, spotlighting possible antitrust and data jurisdiction considerations. The companies will build AI computing infrastructure across Texas and other states.

📈 OpenAI’s rumored $300B valuation reflects sky-high market confidence in AI. Softbank is investing. Throwback to 2018!

🏈 OpenAI put out a Super Bowl commercial. Your editor kind of liked it although interesting that the same CMO that made the Coinbase screen static ad a few years ago was behind it. Can’t beat your editor’s favorite technology ad of all time, Google’s 2009 Parisian Love ad (sniff).

AI & Legal

📱 Our favorite copyright case, NYT v OpenAI, had a big hearing. The transcript hasn’t been released yet, but Bloomberg reporter Aruni Soni says the hearing made clear “NYT and OpenAI disagree over very fundamental aspects of the technology, like whether an AI model ‘knows’ what it is saying.” Watch this space.

🎤 Vanessa Wu, the General Counsel of HR software unicorn Rippling, discussed AI’s potential to streamline contract review and negotiations by embracing AI as a legal team. Most notable? That Rippling has a zero redline policy across 20K customers. Wowza.

⚖️ Another day, another fake case citations instance. This time in Wyoming. Interesting wrinkle that there was local counsel and a national plaintiffs' counsel, and both will be sanctioned. Fortune 10 company Walmart was a litigant. For those of you wondering - this is not a legal AI issue, this is a ChatGPT issue. Regardless, check your cases, folks.

AI Corporate Adoption

🤑 This guide for CFOs outlines AI’s promise in streamlining operations by leading with value.

💪 McKinsey’s concept of “superagency” suggests that empowering employees with AI could be a competitive edge.

⏱️ This AI Proficiency Report by software provider Section reveals that the most effective improvements in productivity (10+ hours saved per week) happen at companies that fully embrace AI. Having an AI policy is not enough, it needs to be an AI culture. In other words, leadership must drive AI, or it doesn’t happen.

“26% of companies [ ] aren’t saying anything [about AI] .… ‘Staying silent’ is directly correlated with [] employees having poor AI proficiency. 35% of AI skeptics and 32% of novices work in a ‘silent company’, compared to 7% of experts.”

AI Globally

🇪🇺 In Paris last week, Vice President Vance told EU regulators to back off on AI regulation. Key quotes:

"AI [will] make us more productive, more prosperous, and more free. The US is the leader in AI, and our administration plans to keep it that way."
"The AI future is not going to be won by hand-wringing about safety. It will be won by building …”
“[W]e must focus now on the opportunity to catch lightning in a bottle … and use AI to improve the well-being of our nations and their peoples."
“[W]e want to embark on the AI revolution in the spirit of openness and regulation. But to create that kind of trust, we need international regulatory regimes that foster the creation of AI technology rather than strangle it.”
“From law to medicine, manufacturing - the most immediate applications of AI almost all involve supplementing, not replacing, the work being done by Americans.” Editors note: we agree deeply on this one!

Full transcript here, or click the image to watch the video.

The AI Action Summit, where Vance spoke, was co-chaired by France and India, gathered global leaders to discuss the development of ethical and sustainable artificial intelligence. The summit emphasized the importance of international cooperation to ensure AI technologies are inclusive, safe, and beneficial for all, while addressing digital divides and promoting sustainable growth.

📵 Senator Josh Hawley proposed a bill that would impose penalties of up to 20 years in prison or $1M in fines for violating its restrictions on AI-related trade and collaboration with China on AI models. Here is GC AI’s summary of the bill.

AI & Intellectual Property

💡 TechCrunch reports on David Sacks's allegations that Deepseek potentially misused OpenAI's models or training data, raising questions about IP rights and ethical boundaries in AI development.

📱🧑‍💼 Nisha Dua shares how 85% of all ChatGPT mobile users are men (!). What we are all asking ourselves, “Where are the women? Pew research found that women are more skeptical of AI versus men; and that 2x as many women wouldn't let their kids use AI versus men.”

New, Notable, Random, & Fun

😂

🎶 Sir Paul McCartney has warned against proposed AI legislation that could allow the use of artists' copyrighted works to train AI models. McCartney says, “The truth is, the money’s going somewhere… Somebody’s getting paid, so why shouldn’t it be the guy who sat down and wrote ‘Yesterday’?”

🫣 Luke de Pulford analyzes DeepSeek’s privacy policy and finds that the collect IP, keystrokes, and device information, and store that data in China. See GC AI analysis of DeepSeek’s privacy policy.