- Anthropic says the transparency-focused AI safety laws it backed in 2025 may already be outdated and is pushing states for tougher rules.
- “Transparency and self-reporting are no longer sufficient safety measures for the most powerful AI systems,” said Anthropic’s head of US state and local government relations, Cesar Fernandez.
- The company has endorsed third-party-auditing bills in Illinois and Massachusetts, the latter empowering the state attorney general to seek injunctive relief.
- Critics including David Sacks call it “regulatory capture”; Anthropic says its supported bills apply only to “large AI model developers” — generally those with more than $500 million in annual revenue.
What Happened
Anthropic, now valued at nearly $1 trillion, is pushing US states to adopt tougher AI regulations, arguing that the transparency-focused safety laws it helped pass in California and New York in 2025 are already outdated, according to a July 16, 2026 Wired interview. “The transparency-focused safety bills of 2025 were a really important start, but as the capabilities of AI systems continue to advance quickly—the policy responses need to match,” said Cesar Fernandez, the company’s head of US state and local government relations.
Why It Matters
A pro-regulation stance is an unusual message from a company that must build a large business selling access to advanced AI, but Anthropic ties it to its founding mission of ensuring a safe transition through transformative AI. As it has grown, the company has begun endorsing some of the nation’s harshest proposed rules on frontier AI, many aimed at catastrophic risks such as models contributing to financial disasters or mass-casualty events. With Congress stalled on AI legislation, states are taking the lead.
Technical Details
Beyond the California and New York self-reporting laws, Anthropic has supported an Illinois measure requiring AI labs to have their safety processes evaluated by third-party auditors, and most recently endorsed a Massachusetts policy that would also require third-party auditing and empower the state attorney general to seek injunctive relief from non-compliant companies. Anthropic says the bills apply only to “large AI model developers” — a term defined differently in each bill but generally covering companies that have spent hundreds of millions on AI development and have more than $500 million in annual revenue. Fernandez, who joined Anthropic this year after leading state government relations for FanDuel and public policy work at Uber, said “it’s hard to imagine a startup meeting that threshold.”
Who’s Affected
The push affects every large frontier lab and the state legislators now setting AI policy in Congress’s absence. It also touches well-funded challengers — Safe Superintelligence, Thinking Machines Lab, and Mistral have each raised billions and could approach the revenue thresholds, though their revenue remains far below Anthropic’s or OpenAI’s. David Sacks, former White House AI czar and a technology adviser to President Trump, has accused Anthropic of “a sophisticated regulatory capture strategy based on fear-mongering” that traps smaller startups in red tape — a characterization Fernandez rejects.
What’s Next
Fernandez frames the goal as inspiring “a race to the top developing the most safe and secure AI systems,” which Anthropic argues requires legislation, not just voluntary commitments. With more state bills expected as Congress remains gridlocked, the open question is whether other large labs follow Anthropic toward third-party auditing or continue to resist, and whether the revenue thresholds hold as billion-dollar startups scale.
Related Reading
- OpenAI Hires for Families as ChatGPT’s Over-35 Audience Grows
- Bridgewater Fine-Tunes Qwen3-235B to Beat GPT and Claude on Finance
- Anthropic Blocks Chinese Firms From Claude Code as Alibaba Bans It Internally
- Researcher Used Claude to Uncover a Ticketing Flaw at Most US Festivals
- Anthropic Launches Claude Science, a Flagship AI Product for Research
