US advances AI safety through state and federal action

Efforts to set safety rules for advanced AI systems are moving forward in the United States at both the state and federal levels, according to OpenAI. The company says California, New York, and Illinois have each taken steps toward a shared baseline for frontier AI governance, with state laws focused on safety frameworks, incident reporting, and independent audits. OpenAI argues that these overlapping state measures could help form a de facto national standard if they continue to align, while also saying that a federal framework is needed for issues that require national security expertise and broader coordination. The company says the Trump administration is working on testing standards for the most capable AI models, especially for cyberrelated risks, and that Congress is also considering related legislation. The post says a national framework would help create consistent rules for evaluating advanced AI systems and support safer deployment for government, critical infrastructure, allies, and other trusted partners. It also says such a framework could lay the groundwork for international AI standards and a USled approach built around democratic values.