OpenAI says the U.S. is advancing AI safety through state and federal actions

OpenAI's Chris Lehane says the U.S. is forming a de facto national standard through frontier AI safety legislation in California, New York, and Illinois, which he calls "reverse federalism." The common elements across the three states are frontier model risk assessment and public disclosure, reporting of serious safety incidents, and independent audits. At the federal level, the Trump administration is developing a government testing framework for the strongest AI models in the cyber domain, with OpenAI participating in related discussions, aiming for implementation by early August.