Illinois passes strongest US AI safety law: mandatory audits, whistleblower protection, and fines up to $3 million
What it really says
On July 6, 2026, Governor JB Pritzker signed the Artificial Intelligence Safety Measures Act (SB 315) into law - described as the strongest AI safety legislation of any US state. The law targets frontier AI models whose developers generate annual revenue exceeding $500 million and whose models are trained with particularly high computing power. Affected companies would include OpenAI, Anthropic, Google, Meta, and xAI. The law contains several core elements: Beginning January 1, 2027, covered companies must file transparency reports before deploying new or substantially modified frontier models. Starting January 1, 2028, annual independent third-party safety audits become mandatory - Illinois is the first US state to require this. Companies must report critical safety incidents within 72 hours and implement a documented risk management framework for preventing catastrophic harm. The Whistleblower Act is amended to protect employees who report AI safety concerns, including through confidential reporting channels. Violations can result in civil penalties of up to $3 million per offense, with the Attorney General holding exclusive enforcement authority.
Our assessment
This story merits a green rating because it represents a concrete and substantial response to concerns that AI development is proceeding unchecked. Together with California's SB-53 and New York's RAISE Act, three US states now cover approximately 40 percent of the US AI market with binding safety regulations - a de facto federal standard without federal legislation. Three aspects are particularly noteworthy: First, the requirement for independent audits goes beyond voluntary self-commitments and creates an objective verification mechanism. Second, the whistleblower protection addresses a central problem - employees at AI companies often have the earliest knowledge of safety risks but cannot report them for fear of professional consequences. Third, the 72-hour reporting requirement for critical safety incidents mirrors existing data protection notification obligations. However, the law applies only to the largest developers and does not take effect until 2027 and 2028 respectively. Smaller AI providers and the deployment of existing models by enterprises remain unregulated.
Relevance for Germany
This law is significant for Germany and Europe for several reasons. First, the US approach is increasingly converging with the European regulatory model. While the EU AI Act already provides a comprehensive framework, the Illinois development shows that even in the US, confidence in voluntary self-regulation is waning - a validation of the European strategy. Second, since the affected companies - OpenAI, Google, Meta, Anthropic - distribute their models globally, they must meet the strictest standards across all markets. The combination of the EU AI Act and US state laws raises the safety standard for models used in Germany. Third, SB 315's whistleblower protection complements the EU Whistleblower Directive and creates a transatlantic protection mechanism for employees reporting safety concerns. Fourth, the 72-hour reporting deadline for safety incidents mirrors the GDPR's notification deadline for data breaches, creating consistent regulatory logic across both sides of the Atlantic.
Fact check
Governor Pritzker's signing of SB 315 on July 6, 2026 is confirmed by the Governor's official press release, Capitol News Illinois, NBC News, and numerous legal publications. The $500 million revenue threshold and compute threshold for covered models are consistently cited across all sources. The two-phase timeline - transparency obligations from January 2027, audit obligations from January 2028 - is confirmed by legal analyses from Crowell & Moring and Buchanan Ingersoll & Rooney. The civil penalty of up to $3 million per violation and exclusive enforcement by the Attorney General are consistently documented by the Transparency Coalition and multiple law firms. The comparison with California (SB-53) and New York (RAISE Act) and the estimate that these three states cover approximately 40 percent of the US AI market comes from Transparency Coalition and NBC News reporting.
Source
- • https://capitolnewsillinois.com/news/pritzker-signs-landmark-ai-regulation-bill-that-aims-to-mitigate-risks/
- • https://www.nbcnews.com/tech/tech-news/illinois-legislature-passes-historic-ai-bill-rcna347191
- • https://www.transparencycoalition.ai/news/illinois-gov-pritzker-signs-landmark-ai-safety-measures-act-into-law
- • https://www.crowell.com/en/insights/client-alerts/illinois-imposes-transparency-and-safety-obligations-on-frontier-ai-systems