Anthropic warns against overriding state AI laws without robust federal rules

• Anthropic says any federal preemption should be contingent on robust national standards, including third-party evaluations of advanced systems.

Anthropic called on the US Congress to avoid overriding state artificial-intelligence regulations unless lawmakers enact a “rigorous” federal framework that tackles “catastrophic AI risks,” and proposed independent safety testing for the most powerful AI models, according to a company statement.

The appeal comes as President Donald Trump has pressed Congress to pass legislation that would supersede state-level AI rules. Anthropic said any federal preemption should be contingent on robust national standards, including third-party evaluations of advanced systems.

Anthropic is preparing for a US initial public offering, a listing that could be among the most consequential in years, with potential implications for index composition and capital flows. The company did not respond to a Reuters request for comment on the timing of its policy proposal.

Advancing Anthropic’s agenda in Washington would likely require support from President Trump. The company’s relationship with the administration frayed earlier this year after Anthropic declined to allow US military use of its AI models for domestic surveillance and fully autonomous weapons systems.

The government responded by placing the firm on a national security blacklist slated to take effect later this year, though tensions have shown signs of easing across parts of the US government as of June 5.

Separately, Anthropic urged Congress and states to modernise unemployment insurance technology ahead of potential AI-driven job losses, warning current systems are not equipped to handle a large labour-market shock.