Illinois lawmakers are weighing new safeguards for advanced artificial intelligence systems under legislation supporters say is needed before the technology outruns government oversight – or humanity.
Senate Bill 315, known as the Artificial Intelligence Safety Measures Act, would require large developers of advanced “frontier” AI systems to publish safety protocols, report critical incidents and undergo independent third-party audits intended to evaluate catastrophic risks tied to the technology.
Scott Weiser of the Secure AI Project described establishing “rules of the road” for AI as one of the most important issues lawmakers could take up this session.
The proposal focuses on large-scale AI systems capable of posing what the bill defines as “catastrophic risk,” including cyberattacks, biological weapons development, major infrastructure threats or autonomous criminal conduct.
During Tuesday’s early morning House Executive Committee hearing, some supporters argued Illinois should begin establishing stricter guardrails since Congress has yet to adopt nationwide AI safety standards.
Industry groups largely did not oppose the bill’s goals but argued proposed audit requirements are premature because no formal infrastructure currently exists for certifying independent AI auditors.'
Representative Bob Morgan, a Deerfield Democrat, questioned opposition to independent verification requirements, while supporters argued similar transparency discussions are already unfolding in states including California and New York.
At one point, one witness warned advanced AI systems are – in fact – already capable of assisting inexperienced users attempting to create biological weapons. Another opponent countered the legislation “requires companies to hire auditors who don’t exist.”
The bill would also further require developers to report certain AI-related safety incidents to state officials and includes whistleblower protections for employees reporting dangerous conduct or violations of the act.
Critics have scrutinized several provisions exempting some AI safety reports and audit materials from public disclosure under Illinois’ Freedom of Information Act – language that supporters say is necessary to protect cybersecurity and trade secrets.
The measure remains under negotiation as lawmakers continue debating how states should regulate rapidly advancing AI systems in the absence of federal standards.
