Illinois House passes bill to require third party AI safety audits

In short: Illinois lawmakers passed a bill that would make major AI labs get independent checks of their safety practices.

What happened

The Illinois House of Representatives passed a bill called SB 315 that targets “frontier” AI labs, which are the companies building some of the most powerful AI systems. Examples mentioned include OpenAI, Anthropic, and Google DeepMind.

If the bill becomes law, these companies would have to hire a third party to audit their AI safety practices. An audit is like an outside inspection, similar to how a restaurant might be checked by a health inspector, instead of only trusting its own internal review.

The bill now goes to Illinois Governor JB Pritzker. Pritzker said on social media that he plans to sign it, and he framed the move as a way to hold large tech companies accountable.

Supporters say the bill matters because it adds an independent step. They argue that today many AI companies mostly report on their own safety work, which can look like “grading your own homework.” The audits could be done by large accounting firms like Deloitte, EY, KPMG, or PwC, or by smaller research groups that focus on testing AI systems.

Not everyone supports SB 315. A tech trade group called Chamber of Progress asked lawmakers to oppose it, arguing it could force companies to share sensitive systems with auditors without clear standards.

Why it matters

Congress has not passed major AI safety laws, so states are filling the gap. For regular people, this could affect how much oversight there is on the AI systems that show up in chatbots, work tools, and other services. It could also shape what future national rules look like, since states sometimes act as a trial run for federal policy.

Source: Wired