OpenAI is publicly supporting a package of California AI bills and calling for mandatory national AI safety rules, according to secondary reporting of a post by Chris Lehane, the company’s chief global affairs officer. OpenAI’s original post was not independently available.
Reuters quoted Lehane as saying, “The United States needs mandatory, capability-based national regulation that can evolve as the technology does.” The Washington Examiner quoted him as writing, “We want to work with Congress on mandatory, capability-based national AI safety regulation,” and, “We've reached a new chapter in AI capabilities, and that demands a new chapter for AI policy.” Those quotations have not been checked against OpenAI’s own page.
According to Reuters, OpenAI said on Wednesday, Sept. 9, 2026, that it was pushing for federal rules out of concern the technology could accelerate its own development. Reuters reported that the company wants Congress to adopt capability-based requirements covering testing, independent assessments, cybersecurity, and incident reporting, and that Lehane said voluntary commitments are not enough.
On the California package, Reuters said OpenAI backed Senate Bill 813, Assembly Bill 1405, Senate Bill 1119, and Assembly Bill 1864, and that the company said it is now supporting some bills it had not previously endorsed. The Examiner said the company supports four California bills that would create a framework for independent AI safety assessments, establish standards for AI auditors, impose new protections against children using AI companion chatbots, and require AI models to follow federal screening standards intended to guard against AI-enabled biological threats. Outlets disagree on the details of that support, so the four-bill endorsement should be read as reported, not independently confirmed.
The outlets also disagree on status. Reuters reported that Gov. Gavin Newsom signed SB 813 and AB 1405 into law on Sept. 9, describing them as establishing a framework for independent third-party evaluation and audits, and that AB 1864 and SB 1119 address screening safeguards against AI-enabled biological threats and chatbot protections for children. The Examiner reported on Sept. 10 that all four bills had passed the California legislature and were headed to Newsom’s desk. Without a matching official California confirmation in hand, that signing timeline remains contested.
OpenAI had already taken a public position on one measure, according to Unite.AI: on Aug. 31, 2026, the company said it supports SB 1119, which would impose safety obligations on operators of companion chatbots used by children, and Ann O’Leary urged Newsom to sign the bill.
A public shift by OpenAI toward binding federal safety rules — and toward some California bills it says it had not previously endorsed — is a business and policy story: it would raise compliance costs for frontier labs and strengthen California’s role as a de facto national standard-setter. Reuters has also noted OpenAI and Anthropic preparing for IPOs in the same coverage climate. The company’s earlier opposition to SB 53, cited in reporting, makes any reversal part of the news. What is settled here is limited; the core claims still rest on secondary accounts with a live conflict over Newsom’s signatures.