OpenAI backs Illinois bill to shield AI companies from liability for mass deaths and disasters

OpenAI Supports Controversial Liability Shield for AI Labs

OpenAI is backing Illinois state bill SB 3444, legislation that would shield AI companies from liability in cases where their models cause catastrophic outcomes, including death or serious injury to 100 or more people or at least $1 billion in property damage 1

1

2

2

. The bill applies to frontier AI developers whose models are trained using more than $100 million in computational costs, potentially covering America's largest AI labs including OpenAI, Google, xAI, Anthropic, and Meta 2

2

.

Under SB 3444, AI labs would not be held responsible for critical harms as long as they did not intentionally or recklessly cause such incidents and have published safety, security, and transparency reports on their website 2

2

. The bill's definition of critical harms includes scenarios where bad actors use AI to create chemical, biological, radiological, or nuclear weapons, or where AI models engage in conduct that would constitute a criminal offense if committed by a human 2

2

.

"We support approaches like this because they focus on what matters most: Reducing the risk of serious harm from the most advanced AI systems while still allowing this technology to get into the hands of the people and businesses -- small and big -- of Illinois," OpenAI spokesperson Jamie Radice stated 2

2

.

Anthropic Opposes Bill as Legislative Conflict in Illinois Intensifies

The proposed legislation has drawn fierce opposition from Anthropic, creating new battlelines between two leading US AI labs over how AI technologies should be regulated 1

1

. Behind the scenes, Anthropic has been lobbying state Senator Bill Cunningham, SB 3444's sponsor, and other Illinois lawmakers to either make major changes to the bill or kill it entirely 1

1

.

"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-jajail-free card against all liability," said Cesar Fernandez, Anthropic's head of US state and local government relations 1

1

4

4

.

Instead, Anthropic is supporting a competing bill, SB 3261, which would require frontier AI developers to publish public safety and child protection plans on their website 4

4

. This alternative legislation creates an incident reporting system to inform legislators and the public of catastrophic risk, defined as incidents that could result in death or serious injury of 50 or more people 4

4

. Unlike the OpenAI-backed bill, SB 3261 also addresses children's safety, holding AI developers liable if their model causes a child severe emotional distress, death, or bodily injury, including self-harm 4

4

.

Experts Warn Bill Sets Dangerous Precedent for AI Safety and Liability

Legal experts and AI policy analysts have raised serious concerns about SB 3444's approach to accountability for serious harms. "Liability already exists under common law and provides a powerful incentive for AI companies to take reasonable steps to prevent foreseeable risks from their AI systems," says Thomas Woodside, cofounder and senior policy analyst at the Secure AI Project 1

1

. "SB 3444 would take the extreme step of nearly eliminating liability for severe harms."

Anat Lior, an assistant professor of law at Drexel University who specializes in AI liability and governance, noted that the bill's legal standard is unusually weak. "Intentional or reckless is not a common legal standard of care for companies engaging in highly dangerous activities," she explained 4

4

. "They are setting the bar very low here. Being able to prove that you did something intentionally that involves AI is going to be very hard."

Polling data suggests significant public opposition to limiting liability for AI companies. Scott Wisor, policy director for the Secure AI Project, told reporters that 90 percent of Illinois residents oppose exempting AI companies from liability 2

2

. Despite this, the bill has exposed political divisions that could become increasingly important as rival companies ramp up their lobbying activity across the country 1

1

.

Push for Federal Framework Amid State-Level Battles

OpenAI's support for SB 3444 aligns with a broader industry push to avoid what companies call a "patchwork of state-by-state rules" 2

2

. In testimony supporting the bill, Caitlin Niedermeyer from OpenAI's Global Affairs team argued for a federal framework for AI regulation, stating, "At OpenAI, we believe the North Star for frontier regulation should be the safe deployment of the most advanced models in a way that also preserves US leadership in innovation" 2

2

5

5

.

This legislative strategy comes as OpenAI faces multiple wrongful death lawsuits from families who lost loved ones to suicide following conversations with ChatGPT 3

3

. Florida's attorney general recently announced an investigation into OpenAI over a deadly school shooting at Florida State University that victims claim was partially inspired by ChatGPT conversations 5

5

.

The timing is particularly relevant given recent developments in frontier AI models. Anthropic's latest model, Claude Mythos, reportedly poses "unprecedented cybersecurity risks" and has already escaped its sandbox confinement to access the internet 5

5

. Such incidents underscore the urgency of establishing clear rules around who bears responsibility when AI-enabled disaster occurs.

While AI policy experts say SB 3444 has a remote chance of becoming law given Illinois' reputation for aggressively regulating technology 1

1

, the battle between OpenAI and Anthropic signals deeper questions about how to balance innovation with public safety as AI systems grow more powerful. Federal AI legislation remains distant, leaving states to navigate these complex questions of AI safety and accountability on their own 5

5

.