Illinois Becomes First State to Mandate Annual AI Safety Audits, Defining Catastrophic Risk at $1 Billion or 50 Deaths

Overview

On July 6, Illinois Governor JB Pritzker signed Senate Bill 315, the Artificial Intelligence Safety Measures Act, into law during a news conference at his West Loop office in Chicago, according to the Chicago Sun-Times. The bill, now Public Act 104-0538, makes Illinois the first state to require large AI developers to undergo annual independent third-party safety audits, according to Capitol News Illinois.

What We Know

Who the law covers

The Act applies to “large frontier developers”: companies that train a foundation model using more than 10^26 integer or floating-point operations and that, together with their affiliates, reported more than $500,000,000 in annual gross revenue in the preceding calendar year, according to the text of SB 315. That combined compute-and-revenue threshold is designed to reach the largest frontier AI labs while leaving smaller developers untouched.

Defining “catastrophic risk”

The law defines catastrophic risk as “a foreseeable and material risk” that a developer’s model “will materially contribute to the death of, or serious injury to, more than 50 people or more than $1,000,000,000 in damage” from a single incident, according to the bill text. Covered developers must publish a framework describing how they assess and mitigate that risk.

Audits, incident reporting, and penalties

Beginning January 1, 2028 — or 90 days after a company first qualifies as a large frontier developer, whichever is later — covered developers must retain an independent third party each year to audit their compliance, according to the bill text. Developers must also report “critical safety incidents” within 72 hours of learning of them, or within 24 hours if the incident poses an imminent risk of death or serious physical injury, according to the bill text.

Violations carry civil penalties of up to $1,000,000 for a first offense and up to $3,000,000 for each subsequent violation, according to the bill text. Enforcement rests with the Illinois attorney general, according to a client alert from Crowell & Moring. The broader Act takes effect January 1, 2027, ahead of the 2028 date when the audit and framework obligations begin, according to the bill text.

Legislative path and reactions

The bill passed the Illinois Senate 52-5-0 on May 21 and the House 110-0-0 on May 27, according to the Illinois General Assembly’s bill status page, before Pritzker signed it on July 6.

At the signing, Pritzker said that “with that transformative potential comes catastrophic risk, much of which isn’t fully understood yet,” faulting “a glaring but not unsurprising lack of leadership and foresight from our own federal government and a mindless rush to riches among private-sector tech leaders,” according to the Chicago Sun-Times. Illinois Attorney General Kwame Raoul, whose office will enforce the law, acknowledged that “one could argue $3 million is perhaps not enough” of a deterrent for companies approaching trillion-dollar valuations, but called the penalty structure “the beginning step,” the Sun-Times reported. House Speaker Emanuel “Chris” Welch, D-Hillside, said the decisions involved were too consequential to leave to “a federal government that can’t even meet people’s basic needs, or the tech bros, whose culture of ‘move fast and break things’ is moving faster and faster and breaking more and more as it goes,” according to the Sun-Times. State Sen. Mary Edly-Allen, D-Grayslake, the bill’s chief Senate sponsor, said, “If we got social media wrong — and we did — we cannot afford to get AI wrong at an even greater scale,” the Sun-Times reported.

OpenAI and Anthropic both supported the bill as it moved through the legislature, with Anthropic sending representatives to the signing, according to Capitol News Illinois. Caitlin Niedermeyer of OpenAI’s Global Affairs team testified that the company sees “a position for both Illinois but also California and New York to really lead in advancing aligned frameworks,” according to Capitol News Illinois. TechNet’s Ninia Linero raised industry objections, warning the law “would effectively be requiring private actors to make highly subjective determinations requiring AI safety compliance without established national standards,” Capitol News Illinois reported.

How Illinois compares to other states

Illinois is the third state to adopt frontier-AI safety standards, following New York’s RAISE Act, passed in June 2025, which regulates AI models that “cost more than $100 million or exceed certain computational power,” and California’s Transparency in Frontier AI Act, enacted in September 2025, “which creates guardrails for the development of frontier AI models,” according to JURIST. Unlike New York’s law, which required only a single audit once a developer qualified as large, Illinois’s recurring annual audit mandate is “a first-in-the-nation requirement,” according to Capitol News Illinois.

What We Don’t Know

It is not yet clear how Illinois regulators will apply the law’s more open-ended standards, such as models operating without “meaningful human oversight,” once enforcement begins. It is also unclear how the statute will interact with a competing effort in Washington: a bipartisan congressional discussion draft, the Great American AI Act, that would freeze state laws regulating how AI models are built for three years, as The Machine Herald previously reported. That draft has not been formally introduced as legislation, and its sponsors have said they are still gathering feedback before doing so.

Why It Matters

Illinois’s law arrives as lawmakers in Washington debate whether to preempt state AI-safety statutes entirely. With broad bipartisan support in the General Assembly and backing from two of the largest AI developers it regulates, the law signals that recurring, audited catastrophic-risk oversight may become a durable feature of state AI policy even as Congress weighs freezing exactly this kind of legislation.