OpenAI Urges US Congress to Enact Mandatory Federal AI Safety Regulations
Artificial intelligence pioneer OpenAI has formally requested that the United States Congress pass mandatory federal legislation governing advanced frontier AI models. In a policy statement published by Chief Global Affairs Officer Chris Lehane, the company argued that rapid advancements in artificial intelligence and the prospective emergence of recursive self-improvement demand enforceable national safety standards rather than voluntary corporate commitments.
OpenAI's proposed regulatory framework recommends capability-based rules that apply to AI systems based on what they can achieve rather than who developed them. Key pillars of the requested legislation include mandatory independent safety testing, standardized third-party audits, enhanced cybersecurity requirements, strict incident reporting protocols, and federally coordinated monitoring for potential systemic risks. The company urged lawmakers to pass legislation before Congress adjourns at the end of the year.
Alongside its federal plea, OpenAI dramatically reversed its position on state-level legislation by endorsing four major California AI safety bills, including Senate Bill 813 and Assembly Bill 1405. These measures establish regulatory pathways for independent model evaluations, standards for AI auditing, protections against AI-enabled biological risks, and safeguards for youth interacting with conversational agents. OpenAI acknowledged that recent unexpected leaps in model capabilities prompted its reconsideration, highlighting the urgent need for comprehensive safety guardrails across the technology landscape.
