Illinois passes nation’s strongest AI safety law, setting baseline for frontier models

The Illinois legislature has passed SB 315, establishing the strictest artificial intelligence safety framework in the United States. The legislation arrives in the wake of President Donald Trump’s cancellation of a federal plan to vet frontier AI models, a move driven by concerns that such oversight might stifle innovation. If signed by Governor J.B. Pritzker, the law will impose rigorous compliance requirements on major AI firms, including mandatory public safety plans, annual reports on independent third-party safety testing, and immediate disclosure of critical incidents.

Under the new regime, companies must report significant safety events to the state within 72 hours, or within 24 hours if there is an imminent risk of death or serious physical harm. The statute also introduces whistleblower protections to encourage employees to report emerging risks that firms might otherwise downplay. To enforce these standards, Illinois will likely rely on the Big Four accounting firms—Deloitte, EY, KPMG, and PwC—to conduct independent audits of AI safety practices, a move aimed at ensuring transparency in a sector where companies currently grade their own homework.

Major industry players, including OpenAI and Anthropic, have endorsed the bill. OpenAI views the legislation as a necessary baseline to prevent a fragmented landscape of conflicting state regulations, while Anthropic notes that the requirements align with safety protocols it already employs voluntarily. Both firms see the law as a mechanism to establish uniform standards that larger developers can meet, potentially creating a higher barrier to entry for smaller competitors lacking the resources for extensive independent auditing.

The bill’s passage highlights the growing divergence between state-level action and federal inaction. Democratic Representative Daniel Didech, the bill’s sponsor, stated that Congress has stalled on meaningful protections, leaving states no choice but to intervene. Didech described the legislation as a testing ground for AI governance, suggesting that Illinois’ approach could provide a roadmap for federal policymakers as public trust in AI technologies remains fragile.

Critics, however, warn of unintended consequences. Adam Kovacevich, CEO of the Chamber of Progress, argued that the law creates a liability-focused regime without clear standards, potentially forcing companies to expose sensitive systems to untested auditors. Despite these objections, Governor Pritzker confirmed his intent to sign the bill, asserting that Illinois is leading the nation in holding Big Tech accountable. The law takes effect on 1 January 2027, carrying civil penalties for violations but excluding private rights of action for individuals.