Anthropic Pushes States to Mandate Third-Party AI Safety Audits, Calling 2025 Transparency Laws Already Outdated
Summary
- • Anthropic is backing tougher state AI laws, saying its own 2025 transparency bills in CA and NY are already insufficient.
- • Company supports mandatory third-party auditing of AI safety in Illinois and Massachusetts, beyond self-reporting.
- • Cesar Fernandez, Anthropic's state policy head: 'Transparency and self-reporting are no longer sufficient safety measures.'
- • Critics including Trump AI adviser David Sacks accuse Anthropic of 'regulatory capture' to disadvantage smaller rivals.
Details
Anthropic backs mandatory third-party AI safety audits
Anthropic supports IL and MA bills requiring AI labs to have safety processes evaluated by independent third-party auditors, not just self-report.
2025 CA and NY transparency laws already deemed outdated
Transparency-focused safety bills passed in California and New York in 2025 are now considered insufficient by Anthropic as AI capabilities advance rapidly.
Bills target companies with $500M+ revenue and massive dev spend
Covered entity threshold generally means companies that spent hundreds of millions on AI development and have $500M+ in annual revenue — effectively only frontier labs.
David Sacks calls it a 'regulatory capture strategy'
Trump's former AI czar publicly accused Anthropic of using fear-mongering to trap smaller startups in red tape and cement its own dominance.
Anthropic has backed AI safety bills in four states
Anthropic has supported AI safety bills in CA, NY, IL, and MA — taking the most pro-regulation stance of any major frontier AI lab.
Massachusetts bill empowers AG with injunctive relief
The Massachusetts proposal would allow the state attorney general to seek injunctive relief against non-compliant AI labs.
Near-$1T company championing its own regulation
Anthropic, valued at ~$965B, argues robust external regulation is essential to safely navigate transformative AI — its founding mission.
Anthropic's state-level AI regulatory push and the policy debate it has sparked.
What This Means
Anthropic's aggressive push for mandatory third-party AI safety auditing at the state level is the most concrete proposal for external AI oversight yet from a major frontier lab. By arguing that even the transparency laws it helped pass in 2025 are already obsolete, Anthropic is signaling that AI capabilities are advancing faster than any regulatory framework can track. This sets up a significant political fault line — Anthropic and safety-aligned groups on one side, pro-innovation voices like David Sacks on the other — debating whether this is genuine safety leadership or strategic market manipulation. Congress's continued inaction means states are filling the vacuum, and Anthropic intends to shape the result.