Illinois Becomes First US State to Mandate AI Safety Audits

Illinois Governor JB Pritzker signed the Artificial Intelligence Safety Measures Act, known as Senate Bill 315, on July 6, 2026. It is the first law in the United States to require the largest AI model developers to undergo annual independent security audits conducted by outside parties.

The law requires so-called large frontier model developers to publish safety frameworks describing how they identify and assess "catastrophic risk," defined as the likelihood of an incident that could cause death or serious injury to more than 50 people, or property damage exceeding $1 million.

Audits set it apart

Unlike similar laws passed earlier in California and New York, Illinois is the first state to mandate annual independent third-party audits. Experts point to this requirement as the most rigorous in the country, since it forces an outside party to verify companies' claims rather than relying on self-reporting.

The law also requires companies to report to state authorities any incidents that could harm Illinois residents within 72 hours of identifying them, a window that shrinks to 24 hours if there is a direct threat to life or health. Enforcement falls to the state attorney general, with penalties reaching $1 million for a first offense and $3 million for subsequent ones.

Backing from industry and lawmakers

Notably, the bill was supported by major AI companies themselves, including OpenAI and Anthropic. Anthropic argued that the new law "takes safety practices that leading labs already follow voluntarily, publishing safety frameworks, transparent incident reporting, whistleblower protections, and helps establish a baseline that every leading AI developer should meet." The bill passed the state general assembly with bipartisan support, facing opposition from only five Republicans in the Senate, and was approved unanimously in the House of Representatives.

We are not going to wait for Congress to act - State Senator Mary Edly-Allen, co-author of the bill

Every breakthrough technology... has delivered benefits while carrying real risks - State Representative Daniel Didech, co-author of the bill

An informal national standard

Illinois joins California and New York, which adopted similar AI safety regulations in 2025. Together, the three states account for roughly 40 percent of the US AI market, effectively creating an informal national standard even in the absence of a unified federal law. Companies operating in these states must now comply with three separate sets of requirements simultaneously while Congress weighs its own regulation.

For Polish companies and developers using models such as Claude or GPT, the key takeaway is that regulatory pressure in the US is rising in parallel with the EU's AI Act, which enters its next enforcement phase on August 2, 2026. Companies operating globally will need to reconcile Illinois's audit requirements with California's and New York's standards, as well as the EU's high-risk provisions, which meaningfully raises the cost of operating across multiple jurisdictions at once.

The law's effective date of January 1, 2028 gives companies more than a year to adjust their internal processes. More US states are likely to follow Illinois's lead before Congress settles on a unified federal regulation covering the entire country.

Sources: Pritzker signs landmark AI regulation bill that aims to mitigate risks (capitolnewsillinois.com), Illinois Sets a New Standard for AI Oversight (governing.com), Gov. Pritzker signs first-in-nation Illinois law requiring third-party safety audits for AI giants (lodinews.com)