OpenAI Backs Illinois Liability Shield Bill

OpenAI is trying to turn AI liability into a compliance framework instead of an open-ended litigation risk, and that is a big shift from simply resisting regulation.

• –The bill could set a precedent for shielding frontier labs from catastrophic-harm lawsuits if they meet reporting and safety-process requirements
• –Its compute-based frontier-model threshold likely sweeps in OpenAI, Google, Anthropic, xAI, and Meta
• –OpenAI is aligning with the broader push for federal preemption, arguing state-by-state rules create fragmentation
• –The politics are rough: the bill appears narrowly tailored to industry, while public support for reducing AI-company liability is likely weak
• –If versions like this advance, the real battleground moves from “can labs be liable?” to “what safety paperwork counts as enough?”