OpenAI publicly endorsed four California artificial intelligence safety bills and urged the implementation of mandatory, capability-based national regulations to govern advanced AI models. The policy push comes amid rapid capability gains and growing scrutiny over recursive self-improvement risks.