Illinois becomes first state to require annual third-party audits for AI companies

Illinois AI Bill Establishes Strictest Safety Standards Yet

Illinois Governor JB Pritzker signed the Artificial Intelligence Safety Measures Act on Monday, positioning the state at the forefront of AI regulation in the United States 1

1

. The landmark legislation, also known as SB 315, introduces transparency and accountability requirements for the largest AI models—those generating more than $500 million in annual revenue and trained using massive computing power 3

3

. "Congress and the president ought to be passing similar legislation, but they've so far been unwilling, because many are captive to special interests that profit from the industry having no regulation," Pritzker stated before signing the bill 4

4

.

First-in-Nation Third-Party Audit of AI Models Requirement

What distinguishes Illinois from California and New York is its groundbreaking mandate for annual third-party audits 5

5

. While New York's version only required a single independent audit when developers became large enough to qualify under the law, Illinois demands ongoing annual evaluations by qualified experts without financial conflicts of interest 1

1

. This provision sparked debate during the General Assembly, with TechNet, a coalition of tech executives, expressing concerns that Illinois would "effectively be requiring private actors to make highly subjective determinations requiring AI safety compliance without established national standards, certifications, or clear regulatory guardrails" 4

4

. The requirements are set to take effect on January 1, 2028 2

2

.

Catastrophic Risk Assessments and Incident Reporting

The new law requires model developers to publish AI frameworks outlining how they identify and assess catastrophic risk, defined as incidents that could cause death or serious injury to more than 50 people or more than $1 million in property damage 1

1

. Developers must report any incidents that could cause harm to the state within 72 hours of identification, or 24 hours if it poses an imminent risk for death or serious physical injury 4

4

. The legislation establishes new reporting standards for the possibility that AI models could be used for large-scale harms, such as providing users assistance in creating chemical, biological or nuclear weapons or committing cyber-attacks 1

1

.

State-Driven Push for AI Regulation Creates De Facto National Standard

Senate sponsor Sen. Mary Edly-Allen emphasized the urgency: "We are not willing to wait for Congress to act" 3

3

. Though Illinois, California, and New York only account for roughly 20% of the national population, lawmakers estimate they represent roughly 40% of the U.S. AI market, effectively creating a de facto national standard 1

1

. The bill passed with broad bipartisan support—only five Republican senators voted against it, and it passed unanimously in the House 4

4

.

OpenAI and Anthropic Back Legislation to Mitigate AI Risks

Both OpenAI and Anthropic supported the Illinois AI bill throughout its legislative process 2

2

. Cesar Fernandez, head of U.S. state and local government relations at Anthropic, called Illinois's pairing of AI transparency requirements and external verification "an important step toward the accountability this technology demands" 5

5

. Anthropic representatives were present at the signing ceremony, and the company had previously referenced its Mythos model as an example of a cyberweapon too powerful to release publicly 1

1

. The legislation also creates a channel for whistleblowers working inside frontier AI labs to safely relay safety concerns without fear of repercussions 2

2

. With federal inaction persisting and Washington stalled on AI-related legislation amid partisan debates, most AI labs now support a national AI safety regulation framework that would eliminate the patchwork of state regulations, but firms like OpenAI and Anthropic are turning to state laws in the meantime 5

5

.