NextFin

Illinois Passes America's Strongest AI Safety Law with Mandatory Third-Party Audits

Summarized by NextFin AI
  • Illinois lawmakers passed a landmark AI safety bill establishing the most stringent state-level regulatory framework in the U.S., requiring independent safety audits for major AI developers.
  • The legislation targets AI companies with annual revenues exceeding $500 million, mandating transparency frameworks and annual third-party audits to ensure compliance.
  • Despite support from major AI firms like OpenAI and Anthropic, the auditing mandate has faced criticism for potentially imposing burdens on smaller companies and startups.
  • The bill's enforcement will be managed exclusively by the Illinois Attorney General, with civil penalties of up to $3 million per violation, reflecting a significant shift in AI governance.

NextFin News - Illinois lawmakers passed a landmark artificial intelligence safety bill on Wednesday, establishing the most stringent state-level regulatory framework in the United States. Senate Bill 315, known as the Artificial Intelligence Safety Measures Act, cleared the Illinois House of Representatives in a unanimous 110-0 vote, following its passage in the state Senate last week. The legislation now heads to Governor J.B. Pritzker, who immediately signaled his intention to sign the measure into law. By requiring the developers of the most powerful AI models to submit to independent, third-party safety audits, Illinois has gone a step further than California and New York, transforming the national regulatory landscape for frontier technology.

The new law targets "frontier" AI developers with annual gross revenues exceeding $500 million, a threshold designed to capture industry giants such as Meta, OpenAI, Google, and Anthropic. Under the legislation, these companies must publish a comprehensive transparency framework detailing how they measure model capabilities, assess catastrophic risks, and respond to safety incidents. Crucially, the bill mandates that these firms retain independent third-party auditors annually to verify compliance. This auditing requirement represents a significant escalation from the self-reporting mechanisms established by California and New York in their respective 2025 legislative sessions.

State Senator Mary Edly-Allen, the Senate sponsor of the bill, described the current state of AI development as the "wild, wild West," arguing that lawmakers could not afford to repeat the hands-off approach initially taken with social media. According to Capitol News Illinois, the legislative push was accelerated by warnings from safety advocates. Scott Wisor, policy director for the nonprofit Secure AI, testified during committee hearings that three leading frontier AI developers had acknowledged in their own safety evaluations that their models could provide meaningful assistance in constructing biological weapons. Anthropic, for instance, chose to withhold its powerful Mythos model from public release, citing severe cybersecurity risks.

The rapid state-level legislative action highlights a growing friction between local governments and Washington. U.S. President Trump issued an executive order in late 2025 aimed at discouraging states from obstructing national AI policy, and has recently floated a second executive order that would direct developers on how to report powerful models to federal officials. However, the lack of comprehensive federal legislation has prompted statehouses to act independently. State Representative Daniel Didech, the House sponsor of the bill, emphasized that while artificial intelligence has the potential to drastically improve global quality of life, such progress is only sustainable if the technology is deployed and developed responsibly.

Surprisingly, the bill received active support from major AI developers, including OpenAI and Anthropic, both of which deployed lobbyists in Springfield. According to a statement from Ann O'Leary, vice president of global policy at OpenAI, the company supports the legislation as part of an effort to establish a de facto national standard. In the absence of federal action, uniform state laws prevent a highly fragmented patchwork of fifty different regulatory regimes. Cesar Fernandez, Anthropic’s head of U.S. state and local government relations, praised Illinois lawmakers for setting a new standard and expressed hope that other states and the federal government would build on this foundation.

Despite the overwhelming legislative support, the auditing mandate has drawn sharp criticism from some industry groups. Ninia Linero, representing the tech executive coalition TechNet, expressed concern during committee hearings that the law requires private actors to make highly subjective safety determinations without established national standards or clear regulatory guardrails. Furthermore, Jeremy Kudon, executive director of the American Innovators Network, warned that the $500 million revenue threshold could eventually sweep in smaller companies and startups as AI capabilities double every few months and compute costs fall. Kudon argued that compliance costs would divert vital capital away from innovation and toward legal and lobbying expenses.

To address some of these concerns, lawmakers adopted amendments that clarified the qualifications for third-party auditors, established protocols to protect proprietary corporate information, and pushed the effective date of the law from 2027 to 2028. The final version of the bill also explicitly rules out a private right of action, meaning individual citizens cannot sue developers for non-compliance. Instead, the Illinois Attorney General holds exclusive enforcement authority, with the power to levy civil penalties of up to $3 million per violation.

Some academic observers view the legislation as a necessary but modest first step. Jonathan Iwry, a fellow at the Wharton Accountable AI Lab at the University of Pennsylvania, noted that even the companies developing these systems are currently unable to reliably predict or control their behavior. Iwry suggested that while transparency and auditing are foundational safeguards, the critical question is whether these rules will serve as a springboard for stronger governance or if they will simply lock in modest baseline standards as the permanent national norm.

Explore more exclusive insights at nextfin.ai.

Insights

What is the origin of Illinois' AI Safety Measures Act?

What are the key technical principles behind AI safety audits?

How does Illinois' AI safety law compare to California and New York's regulations?

What is the current market situation for AI developers under this new law?

How have major AI companies responded to the Illinois AI safety legislation?

What are the latest updates regarding the implementation of the AI Safety Measures Act?

What potential challenges do developers face in complying with the auditing requirements?

What controversies have emerged surrounding the $500 million revenue threshold?

How might the Illinois law impact future AI legislation at the federal level?

What are the potential long-term effects of mandatory third-party audits for AI developers?

What arguments have been made against the auditing mandate by industry groups?

How does the Illinois AI safety law aim to address proprietary corporate information concerns?

What lessons can be drawn from other states' approaches to AI regulation?

What role does the Illinois Attorney General play in enforcing the new AI safety law?

What are the key amendments made to the original AI safety bill?

How does the legislation reflect the current trends in AI development and regulation?

What insights do academic observers provide regarding the effectiveness of the new law?

What are the implications of the law for smaller AI companies and startups?

How might the law influence the relationship between state and federal government on AI policy?

Search
NextFinNextFin
NextFin.Al
No Noise, only Signal.
Open App