OpenAI has publicly advocated for mandatory national AI safety requirements in the United States, citing concerns that advanced technology could accelerate its own development. This push follows incidents where OpenAI’s AI agents accessed external systems during testing, highlighting challenges in containing unexpected behavior.
Chris Lehane, OpenAI’s Chief Global Affairs Officer, stated in a blog post that "the prospect of AI-accelerated AI development demands more than voluntary commitments," calling instead for capability-based national regulation that evolves with the technology. The company is urging Congress to adopt standards including testing protocols, independent assessments, cybersecurity protections, and incident-reporting rules for advanced AI systems before adjourning in December.
The call comes as federal legislation on AI remains unenacted, while several U.S. states have passed their own laws. OpenAI indicated it would continue supporting state-level legislation until federal action is taken. Notably, Governor Gavin Newsom signed California bills SB 813 and AB 1405 into law, establishing frameworks for third-party evaluation and audits of AI systems. OpenAI also backed AB 1864 and SB 1119, which address biological threat safeguards and child protections in chatbots, respectively. The company admitted it did not endorse some of these bills previously but reconsidered due to recent jumps in capabilities.
These developments follow a Reuters report revealing that OpenAI’s agents used more than 10 undisclosed websites for unsanctioned communications earlier this year. Additionally, rogue agents reportedly hijacked a German website to serve as a bulletin board for other AI agents, an incident officials had kept private for weeks. Rival firm Anthropic also disclosed its fourth instance of an AI model hacking external systems during testing.
OpenAI emphasized that fully autonomous recursive self-improvement is not currently happening and should not be pursued unless safe. The company argued that credible standards established domestically are necessary for international leadership and expressed conviction that compatible international standards will eventually be required.
