Anthropic has come out publicly against Illinois Senate Bill 3444, the proposed law backed by OpenAI that would shield AI labs from legal liability if their systems are used to cause mass casualties or more than $1 billion in property damage. In a statement to WIRED, Anthropic's head of US state and local government relations called the bill a "get-out-of-jail-free card" and confirmed the company is actively lobbying Illinois lawmakers to gut or kill it.

This is no longer a philosophical disagreement. The two leading US AI labs are now running opposing lobbying operations in a major state legislature, fighting over the fundamental question of who pays when AI goes catastrophically wrong.

What the Bill Actually Does

SB 3444, sponsored by Illinois senator Bill Cunningham, would grant legal immunity to AI companies if a bad actor uses their model to cause harm, as long as the lab published its own safety framework on its website. That is the entire bar: write your own safety rules, post them publicly, and you are shielded from lawsuits even if your AI is used to create a bioweapon that kills hundreds of people.

OpenAI supports the bill. The company argues it reduces the risk of serious harm while "allowing this technology to get into the hands of the people and businesses of Illinois." OpenAI says it has worked with states like New York and California to create a "harmonized" approach to AI regulation.

Anthropic Is Not Having It

"We are opposed to this bill. Good transparency legislation needs to ensure public safety and accountability for the companies developing this powerful technology, not provide a get-out-of-jail-free card against all liability," said Cesar Fernandez, Anthropic's head of US state and local government relations, in a statement to WIRED.

Behind the scenes, Anthropic has been lobbying Senator Cunningham and other Illinois legislators to fundamentally rewrite the bill. The company's position: frontier AI developers should be held at least partially responsible if their technology enables widespread societal harm. Self-published safety frameworks should not be a liability shield.

Illinois Governor JB Pritzker's office also weighed in, stating that the governor "does not believe big tech companies should ever be given a full shield that evades responsibilities they should have to protect the public interest."

The Real Fight Behind the Fight

Policy experts say SB 3444 has only a remote chance of becoming law. But the real significance is what it reveals about the widening strategic gap between the two most influential AI companies in the US.

OpenAI wants to move fast and regulate later. It is backing bills in multiple states that establish favorable frameworks before Congress acts. Anthropic wants accountability baked in from the start. Both companies are ramping up lobbying operations across the country, and the battles in Illinois, California, and New York are early skirmishes in a much larger war over how America regulates AI.

Thomas Woodside, cofounder of the Secure AI Project, told WIRED that the bill would actually "dismantle existing regulations meant to deter companies from behaving badly." Liability already exists under common law. SB 3444 would remove it.

The bottom line: the company that brands itself as the "safety" lab and the company that brands itself as the "access" lab are now fighting, with real lobbyists and real money, over whether AI companies should be legally responsible when their technology kills people. That question will define the next decade of AI governance.

First reported by WIRED.