On July 6, 2026, Governor JB Pritzker signed Illinois Senate Bill 3312 (the Artificial Intelligence Safety Measures Act) into law. The Act enters into force on January 1, 2027. Illinois is the first U.S. state to mandate annual independent third-party audits of frontier AI developers, following California and New York in enacting state-level frontier AI safety legislation.
The Act defines 'large frontier developer' as an entity that develops a 'frontier model' above a computing resource threshold to be set by implementing regulation. Covered entities must adopt, publish, and annually update a frontier AI framework addressing catastrophic risk assessment (defined as an incident likely to cause death or serious bodily injury to more than 50 persons, or property damage exceeding US$1 million), risk mitigations, governance structures, cybersecurity measures, and independent third-party audit findings. Annual independent third-party audits are mandatory. The Illinois Emergency Management Agency and Office of Homeland Security administer the reporting system, may issue implementing rules, and compile annual reports. The Attorney General enforces civil penalties for noncompliance.
Frontier AI developers with products or users in Illinois must establish enterprise-level AI safety programs, commission annual audits by qualified independent third parties, and build incident reporting pipelines before January 1, 2027. AI deployers and downstream businesses purchasing AI services from frontier developers are not directly regulated by the Act but may face contractual compliance obligations inserted by covered developers. Companies operating across California, New York, and Illinois face cumulative obligations that, in combination, may function as a de facto national AI safety standard given the three states' share of the U.S. AI market.
The computing threshold defining 'large frontier developer' remains to be established by regulation, leaving the practical scope of covered entities open until implementing rules are published. Companies developing large language models or frontier multimodal systems should assess their compute usage against likely threshold ranges and prepare for coverage as a precautionary step. The Act does not impose obligations on AI deployers or consumers who are not themselves frontier developers.
Licentium advises AI companies on U.S. state-level AI regulatory compliance and helps clients map obligations across Illinois, California, New York, and EU jurisdictions. Our partner network includes U.S. counsel experienced in state AI legislation and frontier AI governance. Work we undertake includes AI safety program design, frontier developer compliance, third-party audit preparation, incident reporting system design, and multi-state AI regulatory analysis.
Source: Illinois General Assembly, SB3312 (Artificial Intelligence Safety Measures Act), signed July 6, 2026