Anthropic has publicly endorsed California’s SB 53, a bill that could establish the first broad legal requirements for advanced AI companies in the United States.
Proposed by Sen. Scott Wiener, the bill would require major AI developers to publish safety guidelines, share risk assessments, and strengthen whistleblower protections. The legislation aims to make voluntary AI safety commitments mandatory and create a state-level reporting system for critical AI incidents.
SB 53 focuses on transparency, requiring companies with annual revenues over $500 million and operating large-scale models to disclose how they mitigate potential risks such as cyberattacks and misuse of powerful AI tools.
The bill has strong support in California’s Assembly and Senate and is expected to pass by the end of the week. Industry groups, however, have criticized the measure, warning it could slow innovation and push investment to other states.
Anthropic said it backed SB 53 after reviewing lessons from California’s earlier SB 1047, which was vetoed last year by Gov. Gavin Newsom over concerns it would stifle development. SB 53 narrows its focus to transparency and verification, avoiding the liability-heavy provisions of the earlier bill.
If signed into law, SB 53 would make California a national leader in AI regulation, with potential ripple effects across the country and globally.
