Illinois Senate Advances AI Regulation Bill, Joining California and New York
The Illinois Senate voted 52-5 on Thursday to advance Senate Bill 315, which would require large artificial intelligence developers to adopt transparency frameworks, employ third-party auditors, and report on catastrophic risk capabilities. The bill targets companies like Meta, OpenAI, and Anthropic with revenues exceeding $500 million.
The measure is part of an eight-bill package and mirrors legislation passed in California and New York in late 2025. State lawmakers are attempting to establish what they call a "de facto" national AI for Government standard, citing the lack of federal action as their reason for moving forward with state-level rules.
What the Bill Requires
Companies would need to create and publish transparency frameworks explaining how they measure model capabilities, assess catastrophic risk, and respond to safety incidents. The bill defines "frontier" AI as large, groundbreaking systems developed by leading tech companies.
State Sen. Mary Edly-Allen's bill goes further than its California and New York counterparts by requiring third-party audits to ensure compliance. Amendments passed this week clarify who can conduct audits, what they must cover, and how to protect proprietary information.
The effective deadline was extended from 2027 to 2028. Companies must file disclosure statements and pay fees under the revised language.
Industry and Safety Concerns Collide
OpenAI and Anthropic testified in support, arguing the bill protects consumers while allowing smaller companies to innovate. Both companies emphasized the focus on the "most capable" models and worst-case harms.
Three leading frontier AI companies reported in their own safety evaluations that their models could assist in building biological weapons, according to testimony from Scott Wisor, policy director for Secure AI. Anthropic said one of its models is so powerful the company cannot release it publicly.
Startup advocates raised different objections. Jeremy Kudon, executive director of the American Innovators Network, called third-party audits an "expensive requirement" that forces small companies to divert resources toward legal compliance rather than innovation.
Kudon also warned that Illinois's auditing requirement could undermine the national standard three states are trying to establish. "This kind of violates the principle that we set out at the very beginning," he said.
The Definition Problem
Opponents flagged another issue: the bill's definition of "frontier" models may capture more companies as AI technology evolves. Smaller firms working with open-source models or adapting existing systems could fall under the rules unintentionally.
The $500 million revenue threshold may also become less meaningful. As AI becomes more prevalent and companies grow faster, that revenue figure represents less market power than it does today. Model capabilities double roughly every four months, Wisor said.
Jonathan Iwry, a fellow at the Wharton Accountable AI Lab at the University of Pennsylvania, called the bill a modest baseline. "Even the companies developing them seem unable to reliably predict or control the systems' behavior," Iwry said. "Against that backdrop, requirements like transparency, testing and auditing should probably be understood as foundational safeguards rather than the upper limit of accountability."
What Comes Next
The bill now moves to the Illinois House. Its passage reflects what Iwry described as "a practical decision to pursue incremental progress and build on approaches that lawmakers believed were politically achievable."
The real test, he said, is whether these laws become a springboard for stronger AI governance or settle in as the long-term baseline. The ecosystem of AI stakeholders-companies, startups, safety advocates, and regulators-have diverging interests that will shape how rules develop.
Meanwhile, President Trump has signaled he may issue a second executive order directing developers on how to report powerful models to federal officials. His previous order discouraged states from regulating AI, but state lawmakers proceeded anyway, citing federal inaction.
Your membership also unlocks:
