Key facts
- OpenAI is pushing for mandatory national AI safety requirements.
- The company will support state legislation until Congress acts.
- OpenAI supports four California bills: SB 813, AB 1405, SB 1119, and AB 1864.
- The California bills address safety assessments, auditor standards, youth protections, and safeguards against AI-enabled biological threats.
OpenAI announced its support for mandatory national AI safety requirements in the United States, stating it will continue to back state-level AI legislation until federal action is taken. The company detailed its support for four specific bills in California: SB 813, which focuses on infrastructure for independent safety assessments; AB 1405, concerning AI-auditor standards; SB 1119, aimed at protecting young people; and AB 1864, which addresses safeguards against AI-enabled biological threats.
In a blog post, OpenAI emphasized the need for meaningful action on AI policy, advocating for capability-based national AI safety regulation. The company also plans to advance industry-led standards and advocate for compatible international approaches to managing AI risks. OpenAI's Chief Scientist, Jakub Pachocki, has previously written about the necessity of "extreme caution" due to the rapid advancement of AI capabilities, including the potential for recursive self-improvement. Greg Brockman has also highlighted a "defenders window" for establishing safeguards before powerful AI offensive capabilities become widespread.