Illinois requires independent audits of frontier AI models

Accountability reached artificial intelligence. Illinois Governor J.B. Pritzker signed the AI Safety Measures Law (S.B. 315), making the state the first to require annual independent audits for frontier AI models.

Who it applies to

The law reaches frontier model developers: companies with revenues above US$500 million that train advanced AI systems. Violators face fines of up to US$3 million.

The goal

Third-party audits seek to verify that models are trained and deployed with verifiable security controls, transparency and risk management, instead of the voluntary self-regulation that has dominated so far.

The domino effect

With Illinois, the US regulatory trend — until now fragmented among states like Colorado — consolidates. For big tech companies, compliance becomes a cost variable, and for the rest of the world, a model to follow in AI oversight.

The key figure

US$3 million fine. The price of not auditing the most powerful AI models on the planet.