Senior U.S. officials are in active discussions with the chief executives of leading AI companies about taking federal equity stakes in their firms, with OpenAI CEO Sam Altman among those pitching the arrangement, NOTUS reported. Discussions have centered on having firms voluntarily cede shares to the government, with returns potentially directed to public purposes such as a dividend payment to American households. A person familiar with the matter said Anthropic is not having those conversations with the administration. The reporting follows Sen. Bernie Sanders' June 1 preview of the American AI Sovereign Wealth Fund Act, which would take a 50% public stake in the largest American AI firms through a one-time stock tax.
Read at NOTUS ↗
Anthropic on Thursday published a blog post warning that frontier AI systems are approaching recursive self-improvement, the point at which AI can autonomously design and build its own successors. The company cautioned that small misalignmentsAlignmentThe problem of making an AI system reliably pursue the goals its developers and users intend, and the research field devoted to it. Misalignment covers everything from a chatbot flattering users to a capable system deceiving or resisting its operators. in today's models could compound across AI-built generations, growing more frequent but less understood until humans lose control, with development proceeding at the speed of computeComputeThe processing power used to train and run AI models, usually measured in chips, GPU-hours or FLOPs. rather than human oversight. Anthropic called on frontier developers to establish a verifiable, coordinated way to slow or temporarily pause development if risks rise, and said more than 80% of the code merged into its codebase as of May 2026 was generated by Claude. The Anthropic Institute will research the verification systems needed to support a credible global slowdown.
Read at Wall Street Journal ↗ • Read at Reuters ↗ • Read at Anthropic ↗
Reps. Jay Obernolte (R-Calif.) and Lori Trahan (D-Mass.) unveiled a 269-page discussion draft Thursday that would preemptPreemptionWhen federal law overrides state law on the same subject. In AI policy it refers to proposals, in Congress or by executive action, to block or freeze the growing patchwork of state AI laws in favor of a single national standard. state laws regulating frontier AI development for three years, require developers with more than $500 million in annual revenue to publish safety frameworks and retain NIST-licensed independent auditors, and establish the Center for AI Standards and InnovationCAISIThe Center for AI Standards and Innovation, the US government's AI testing body inside NIST at the Commerce Department. Formerly the US AI Safety Institute, it evaluates frontier models with developers on a voluntary basis and works with the UK's AI Security Institute. in statute with $100 million a year through fiscal 2029, Roll Call reported. Developers could face penalties of up to $1 million per day for safety or audit violations. The draft drew immediate pushback over preemption — including from the three House Democrats tapped by Minority Leader Hakeem Jeffries to lead on AI, who said it "cannot serve as the basis for productive dialogue" — and faces Republican leadership skepticism that most likely pushes any serious action to 2027, Politico reported.
Read at Roll Call ↗ • Read at Politico ↗
The House Armed Services Committee voted out the fiscal 2027 National Defense Authorization Act late Thursday after 14 hours of debate covering roughly 900 amendments, The Hill reported. The $1.15 trillion bill heads to the House floor with AI-related amendments covering report language on AI adoption and deployment, AI-enabled advanced manufacturing, and requiring human oversight in AI-testing and evaluation. Notable failed amendments include language addressing lethal autonomy and human-in-the-loop requirements.
Read at The Hill ↗ • Read at House Armed Services ↗