Overview
- OpenAI endorsed Illinois SB 3444, which would protect developers of advanced AI systems from lawsuits over defined “critical harms.”
- The bill defines a frontier model as one trained with more than $100 million in computing costs, a threshold that likely covers OpenAI, Google, xAI, Anthropic, and Meta.
- Critical harms include events like death or serious injury to at least 100 people, $1 billion or more in property damage, and AI assistance in creating chemical, biological, radiological, or nuclear weapons.
- The liability shield would apply only if a lab did not act intentionally or recklessly and posted safety, security, and transparency reports, and OpenAI’s Caitlin Niedermeyer urged a federal framework since Congress has not set liability rules.
- Policy experts in Illinois say the measure faces long odds and cite polling that 90% oppose exempting AI firms, as families suing OpenAI over alleged ChatGPT-linked suicides pursue separate cases.