Illinois Gov. JB Pritzker signed state Senate Bill 315 (SB 315) into law Monday, establishing sweeping bipartisan safety and accountability measures to guide responsible advancement of frontier AI models.
The Illinois legislature passed SB 315 in May. Although the federal government has attempted to rein in state-level AI regulation, states like Colorado and California have enacted their own laws. The U.S. Congress, meanwhile, has yet to pass a comprehensive federal AI law.
Even the federal government, though, has indicated its recognition of the significant risks posed by frontier AI models — the most advanced AI models, which present significant security risks — and that is the focus of SB 315. The bill passed unanimously in the Illinois House of Representatives on May 27 and by a vote of 52-5 in the Illinois Senate the week prior. The law will take effect Jan. 1.
“There is a reason that this legislation earned overwhelming and bipartisan support in the General Assembly and has widespread public support,” Gov. JB Pritzker said Monday during a press conference, during which he signed the bill into law. “People want protection from the risks of AI, and companies want the regulatory certainty to continue developing the technology.”
The new law does several key things. It requires large AI companies to develop risk mitigation frameworks and undergo mandatory, annual, independent third-party audits — a first for any state AI legislation, per Pritzker. Developers are also compelled to report critical safety incidents within 72 hours, or 24 hours if the incident creates a risk of death or serious physical harm. The law also establishes protections for whistleblowers, creating confidential channels for employees to report safety risks.
Companies should not be able to continue advancing AI technology without guardrails in place, as underlined by Thomas Woodside, co-founder and policy adviser at the Secure AI Project: “But passing a law like this can be very difficult in the face of immense money and power.”
Pritzker critiqued the lack of parallel action by Congress and President Donald Trump, noting that many policymakers are unwilling to do so because of the influence of special interests. Guardrails, the governor said, are not intended to stifle AI technology’s potential — in fact, quite the opposite. The governor stated a goal which mirrors that of the federal government: to help position the U.S. as a world leader in this area.
With historical technology transformations, like electricity and air travel, the government responded by building safeguards, state Rep. Daniel Didech, one of the bill’s sponsors, said, indicating industry partners like Anthropic have supported this legislation.
“When we required seatbelts and airbags, and utilized crash test dummies, and built literal guardrails on our highways, it didn’t stifle automobile innovation; it made it safer for cars to go faster,” Didech said.
The legislation targets the largest and most advanced frontier AI developers and systems capable of writing code, generating synthetic media, conducting complex reasoning, and “influencing millions of people at scale,” another bill sponsor, state Sen. Mary Edly-Allen, said during the conference. She noted that these technologies pose risks related to biological threats, mass misinformation campaigns, fraud, election interference and cyber threats.
“SB 315 received strong bipartisan support because it’s not a partisan issue,” Edly-Allen said. “AI will continue to affect every community, every business, and every family.”
Leave a Reply