On the morning of July 6, at a signing ceremony in Chicago, Illinois Governor JB Pritzker put his name on a law no other state has attempted. The Artificial Intelligence Safety Measures Act requires the companies building the largest AI models to open their safety practices to independent outside auditors, every year, whether they like it or not.
The law, which moved through Springfield as SB 315, makes Illinois the third state to regulate so-called frontier models, the biggest and most expensive AI systems on the market. New York got there first with its RAISE Act, and California followed with SB 53, the Transparency in Frontier AI Act. Illinois borrowed from both, then went further. Neither of the earlier laws requires an independent third-party audit. SB 315 does, and that single provision is why policy watchers are calling it the strongest AI safety statute in the country.
What the law actually requires
The core obligations land on large frontier developers, the handful of companies training models at the outer edge of scale and cost. Each of them must create, implement, publish and annually update a frontier AI framework covering catastrophic-risk assessment, risk mitigations, cybersecurity, internal governance, third-party evaluations and the risks that arise when a company turns its own models loose inside its own operations.
Before deploying a new or substantially modified frontier model, a developer must publish a transparency report along with a summary of its catastrophic-risk assessment. Then comes the piece that sets Illinois apart: an annual independent audit of the company’s safety practices, with rules governing auditor access, reporting, record retention and publication of results. No state has demanded that before.
Developers must also report critical safety incidents and submit periodic summaries of internal-use risk assessments. The Illinois Emergency Management Agency and Office of Homeland Security will administer the reporting machinery in consultation with the state attorney general. The law protects whistleblowers inside AI companies, sets civil penalties for violations, and explicitly declines to create a private right of action, meaning enforcement stays with the state rather than the plaintiffs’ bar.
Written with industry at the table
The bill’s sponsors, Senator Mary Edly-Allen of Grayslake and Representative Daniel Didech of Buffalo Grove, both Democrats, steered it to passage with what they describe as overwhelming bipartisan support. Edly-Allen said the law is about “setting clear expectations for transparency, accountability and public safety” before a preventable catastrophe occurs, not after one.
The drafting process was unusually broad. Edly-Allen said her team worked with more than 150 stakeholders, advocates and industry players, a list that included the Secure AI Project, the Transparency Coalition and the AI company Anthropic. Steve Wimmer, a senior advisor at the Transparency Coalition who worked with legislators through the session, called the result “a new template for responsible AI governance” and predicted that other states, and eventually Congress, will follow it.
The states are not waiting for Washington
Illinois is one data point in a much larger pattern. Rhode Island enacted three AI laws in late June, including a ban on AI therapy services offered without a licensed professional and a self-harm safety requirement for chatbot operators. New York’s legislature closed its session by sending Governor Kathy Hochul a stack of AI bills, among them a training-data transparency act and a one-year moratorium on permits for hyperscale data centers drawing more than 20 megawatts. The Transparency Coalition counted 78 chatbot safety bills alive in 27 states just six weeks into the 2026 legislative season.
Missouri, meanwhile, passed a measure that would fine companies $10,000 for advertising an AI chatbot as capable of providing therapy or a mental health diagnosis, with the penalty doubling for repeat offenses. It awaits the governor’s signature. Massachusetts legislators spent last week fighting over the final shape of a comprehensive privacy bill that would make it the twenty-first state with such a law.
The open question now is how the frontier labs respond. Complying with one state’s audit regime is manageable. Complying with three different ones, drafted by three different legislatures, is the kind of patchwork that historically drives industries to ask Congress for a single federal standard. Whether Washington obliges may decide if the Illinois model becomes the national one. For more coverage of AI policy and regulation, visit Mylistingo.
What the Illinois AI Audit Law Requires
The Illinois Artificial Intelligence Accountability and Transparency Act, signed into law in June 2026, requires any organization deploying high-risk AI systems that make or significantly influence consequential decisions about Illinois residents to conduct independent annual audits. High-risk systems include algorithms used for employment decisions, credit scoring, housing eligibility, healthcare access, criminal justice assessments, and insurance underwriting. The audits must evaluate the AI system for bias, accuracy, transparency, and compliance with existing anti-discrimination laws. Audit results must be submitted to a newly created Illinois AI Regulatory Office, which will maintain a public registry of audited systems and their compliance status.
Potential Impact on AI Regulation Nationwide
Illinois has a history of pioneering technology regulation that other states later adopted — the state’s Biometric Information Privacy Act (BIPA) served as a model for similar laws in Texas, Washington, and California. Legal experts expect the AI audit law to follow a similar trajectory, with at least a dozen states considering comparable legislation. The law’s independent audit requirement is particularly notable because it creates a role for third-party auditors similar to the financial auditing industry. This could spawn an entirely new profession of AI auditors, with certification programs, professional standards, and liability frameworks for auditors who certify systems that are later found to be discriminatory or non-compliant.
Industry Response and Compliance Challenges
The technology industry has responded to the Illinois law with a mix of support and concern. Major AI companies including Microsoft, Google, and OpenAI have publicly supported the principle of AI accountability while raising practical concerns about the cost and complexity of compliance. Small and medium-sized businesses that use AI tools may struggle to afford independent audits, potentially limiting their ability to deploy AI systems in regulated domains. The law includes provisions for a small business hardship waiver and permits group audits for standardized AI systems used across multiple organizations, but critics argue these accommodations may not go far enough to prevent the law from disproportionately burdening smaller players.
Related: Feds order robotaxi companies to stop blocking first responders
