AI Safety Measures Beginning, Starting in Illinois

Illinois has just enacted the “Artificial Intelligence Safety Measures Act (AISMA), a first-of-its-kind act in the US that requires annual third-party audits of frontier AI models.

Frontier AI models are defined as the most advanced, general-purpose artificial intelligence systems available, capable of performing a wide range of tasks such as reasoning, multimodal understanding, and autonomous task execution.

They are designed to adapt to various applications and require rigorous governance due to their powerful capabilities. To comply with this act, developers must conduct catastrophic risk assessments, provide transparency reports, report critical safety incidents and submit periodic reports to regulators.

This law goes a step beyond similar laws in CA, NY and the EU by requiring the independent audits.