OpenAI and Anthropic Split Over Massachusetts AI Safety Rules

OpenAI has endorsed Massachusetts Senate legislation that would establish new safety requirements for AI developers, positioning itself against a more stringent House bill backed by its competitor Anthropic.

The Senate proposal would mandate that leading AI companies publish safety frameworks and catastrophic risk assessments. OpenAI spokesperson Liz Bourgeois said the company is urging lawmakers to add mandatory third-party audits to the bill, which would align it with recently passed Illinois legislation.

Even with those audits included, the Senate version would remain less demanding than the House bill Anthropic endorsed last month. That legislation requires independent safety evaluations—considered a higher standard than third-party audits—and grants expanded enforcement authority to the Massachusetts attorney general.

Why it matters

The competing endorsements reveal a fundamental divide in how major AI companies want to be regulated. OpenAI is pursuing what it calls "reverse federalism," encouraging states to adopt similar baseline rules that could form a national standard. Anthropic is pushing states to implement progressively stronger safeguards, potentially creating a patchwork of varying requirements across the country. Massachusetts lawmakers now face pressure from both sides as they reconcile the two bills.

Diverging regulatory strategies

Cesar Fernandez, Anthropic's head of U.S. state and local government relations, described the Massachusetts House bill as the strongest AI safety proposal in the nation during an interview with POLITICO last week. He drew a clear distinction between the two companies' lobbying approaches: while OpenAI works to harmonize state rules with existing laws in places like Illinois, California, and New York, Anthropic advocates for increasingly rigorous state-level regulations.

Bourgeois said OpenAI only began engaging with Massachusetts legislators within the past week and that the company's lobbyists are currently in discussions with the Senate bill's sponsor about incorporating audit requirements.

Federal vacuum drives state action

OpenAI framed its support for the Massachusetts Senate bill as part of a broader strategy to address the absence of federal AI regulation. "In the absence of federal action, we support state efforts like this one in Massachusetts — alongside legislation already in place in California, New York and Illinois — helping to create a de facto nationwide approach," Bourgeois said.

The divergent positions are likely to intensify lobbying by both companies as Massachusetts legislators work to finalize AI safety legislation. The outcome could influence how other states approach AI regulation and whether a uniform national framework emerges from state-level initiatives.

These details were first reported by POLITICO.