Illinois House Approves Frontier AI Safety Bill Mandating Third-Party Audits
SB 315 represents the strongest artificial intelligence safety regulation in the United States, shifting oversight from self-regulation to mandatory external audits by firms such as Deloitte, EY, KPMG, or PwC.
The Illinois House of Representatives has passed Senate Bill 315, legislation that mandates independent third-party audits for frontier artificial intelligence laboratories. The bill requires companies including OpenAI, Anthropic, and Google DeepMind to have their safety practices verified by an external body to ensure adherence to their own stated safety standards. Governor JB Pritzker announced via social media that he intends to sign the bill into law, citing the necessity of holding Big Tech accountable for its operations.
If enacted, the legislation is described by AI safety experts as the strongest of its kind in the United States. It moves beyond self-regulation by introducing mandatory external verification of safety practices. Scott Wisor, policy director at the Secure AI Project, noted that the bill addresses a situation where AI companies currently grade their own homework, requiring an independent auditor to check whether labs adhere to their safety commitments.
Under the proposed framework, AI laboratories may utilise major accounting and auditing firms, specifically Deloitte, EY, KPMG, and PwC, to conduct these audits. Alternatively, labs may engage members of the AI Evaluator Forum, a coalition of smaller research organisations including METR, Transluce, and Averi, to assess adherence to safety standards. Illinois Representative Daniel Didech, a sponsor of the bill, stated that state-level legislation serves as a testing ground for potential future federal laws.
Industry reaction has been divided. OpenAI has formally endorsed SB 315, with its chief of global affairs, Chris Lehane, stating that the company’s AI policy is now oriented around passing similar state laws. Anthropic claims to have been the first AI lab to support the bill, with its head of US state and local government relations, Cesar Fernandez, stating the legislation will help establish a baseline for leading developers.
Conversely, the Chamber of Progress, a trade group representing companies such as Google, Apple, and Amazon, has sent a letter to Illinois lawmakers urging them to oppose the bill. The group argues that the legislation would force companies to expose sensitive systems to untested auditors. This development follows a shift in OpenAI’s position, as the company previously supported a different Illinois bill that included a liability shield for catastrophic harm, a stance Lehane later described as an oversight.
