Illinois becomes first state to require third-party audit of AI models

Illinois Gov. JB Pritzker (D) on Monday signed a bill making the state the country’s first to mandate that the largest artificial intelligence labs obtain third-party audits of their safety plans.

Pritzker signed the Artificial Intelligence Safety Measures Act, or S.B. 315, in a ceremony Monday, stating people “want protection from the risks of AI.”

The bill mirrors similar legislation in California and New York but goes a step further by mandating frontier labs with more than $500 million in revenue to submit third-party audits of their safety plans every year.

The third-party evaluations are intended to secure oversight of these AI firms “by qualified experts without financial conflicts of interest,” Pritzker added.

Like the legislation in California and New York, the bill also requires large frontier developers to create, publish and annually update an AI framework with assessments on catastrophic risks, cybersecurity and more.

The bill passed the state Legislature earlier this year with bipartisan support and received endorsements from ChatGPT maker OpenAI and Anthropic, which created the landmark Claude model.

“As AI systems become more powerful and the federal government is unwilling to step in, states have a responsibility to protect our people from the dangers of AI while still harnessing the unique potential of the technology,” Pritzker said in a press release Monday.

Washington has stalled on getting most AI-related legislation across the finish line amid partisan and intraparty debates.

Most AI labs support a national safety framework for AI that would eliminate the patchwork of state regulations, but with inaction in Congress, firms like OpenAI and Anthropic are turning to state laws in the meantime.

A proposed bipartisan framework on AI, which would include some preemption of state laws, was floated in the House last month, but pushback still remains.

Cesar Fernandez, head of U.S. state and local government relations at Anthropic, said Illinois’s pairing of AI transparency requirements and external verification is “an important step toward the accountability this technology demands.”