OpenAI sends letter urging Newsom to weaken California SB 53

Chris Lehane, OpenAI’s head of global affairs, sent an open letter to Governor Gavin Newsom arguing against SB 53, California’s frontier-AI transparency bill, a year after Newsom vetoed the state’s previous, broader attempt at AI safety legislation. SB 53 had already been narrowed from Senator Scott Wiener’s earlier draft, and unlike the vetoed SB 1047 it did not include shutdown-capability requirements or criminal liability provisions.

OpenAI’s letter did not ask Newsom to veto the bill outright. Instead it proposed that developers who complied with a comparable federal or international framework — naming the EU AI Act’s voluntary Code of Practice, or a “safety-oriented agreement” with a US federal agency — should be treated as automatically compliant with California’s requirements. Critics, including AI policy researcher Miles Brundage, characterised the ask as an attempt to reduce the practical force of state-level oversight by letting compliance with a looser standard substitute for California’s own, rather than a genuine harmonisation proposal.

The letter set OpenAI apart from Anthropic, which had already indicated support for the bill and went on to publicly endorse the version Newsom eventually signed. Newsom signed SB 53 into law on 29 September 2025, without adopting OpenAI’s proposed carve-out, making the letter a marker of the industry’s divided lobbying position rather than a successful intervention — OpenAI got a bill it had opposed passed largely on its original terms, while Anthropic’s more accommodating stance left it aligned with the outcome.