Illinois' House of Representatives just passed a local bill that would require top AI companies to submit model safety plans, submit to third-party audits, and establish whistleblower protections. That makes Illinois the third state — after California and New York — to push forward its own AI safety legislation. The federal government, meanwhile, is heading in the opposite direction. President Trump has signaled he wants to block state-level AI oversight, arguing that pre-approval requirements would undermine the US's competitive edge over China.
https://twitter.com/OpenAINewsroom/status/2059833892316856598
Trump has indefinitely delayed signing an executive order on large model reviews, and even pulled a draft safety review altogether hours before it was supposed to be signed. The White House's original draft would have required developers to share models with the government up to 90 days before public release. To prevent state laws from getting in the way, the White House appointed former Attorney General Pam Bondi to the President's Council of Advisors on Science and Technology (PCAST) on May 26, signaling it's prepared to use litigation to block local regulations that conflict with federal policy.
Industry groups like the Chamber of Progress oppose the local bill, arguing the audit system isn't mature enough yet. The Computer & Communications Industry Association (CCIA) also points out that there's no standard ecosystem for independent audits. But OpenAI and Anthropic have both publicly expressed support for Illinois' auditing bill.