Key facts
- Anthropic is advocating for states to implement stricter AI regulations, including independent audits and safety plan reviews.
- This approach contrasts with OpenAI's strategy of promoting uniform, less stringent state-level AI rules.
- Anthropic's head of U.S. state and local government relations, Cesar Fernandez, stated the company seeks to 'meaningfully raise the bar on safety for the most capable AI systems'.
- The company supported California's 2025 AI regulation law and has backed stricter measures in New York, Illinois, and Massachusetts.
- An Illinois law mandates annual independent third-party audits of AI safety plans for leading companies.
- Anthropic is backing political candidates whose views on AI safety align with its mission.
Artificial intelligence firm Anthropic is pursuing a strategy of encouraging states to enact increasingly stringent AI regulations, including independent audits and safety plan reviews. This approach contrasts with its rival, OpenAI, which has advocated for more uniform, less restrictive rules across states, a tactic termed 'reverse federalism'.
Cesar Fernandez, Anthropic's head of U.S. state and local government relations, stated that the company is not seeking to establish a ceiling for federal legislation but rather to 'meaningfully raise the bar on safety for the most capable AI systems.' He noted that while a federal framework is desired, action on AI risks cannot wait for Washington.
Anthropic was the sole leading AI lab to endorse California's 2025 AI regulation law and has since supported more ambitious bills in New York, Illinois, and Massachusetts. Fernandez indicated that transparency and self-reporting are no longer considered sufficient, citing Anthropic's powerful Claude Mythos model's capability to exploit security flaws.
An Illinois law, signed this month, mandates annual independent third-party audits of safety plans for leading AI companies. Anthropic has also backed proposed regulations in Massachusetts that include hiring independent evaluators for catastrophic risks and empowering the state's attorney general for enforcement.
OpenAI spokesperson Liz Bourgeois defended its 'reverse federalism' approach, arguing it helps regulators enforce laws, provides clearer public protections, and allows developers to focus on safety over conflicting requirements. The company is reviewing the Massachusetts proposal but supports the focus on safeguards.
The two companies have also clashed politically, with dueling super PAC networks funded by each, and Anthropic directly contributing to California legislators whose views on AI safety align with its mission.