OpenAI has announced its support for an Illinois state bill that would protect artificial intelligence developers from legal liability in catastrophic scenarios involving their AI systems. The legislation, known as SB 3444, would shield AI labs from responsibility when frontier models cause extreme harms—defined as deaths or injuries to 100 or more people or at least $1 billion in property damage—provided the developers did not intentionally or recklessly cause the incident.
OpenAI Seeks Protection From AI Catastrophe Liability
The move represents a notable shift in OpenAI's legislative approach. While the company has previously taken defensive positions on AI regulation, this initiative marks a more proactive stance. The bill specifically targets frontier AI models, classified as systems requiring more than $100 million in computational costs to develop. This definition likely encompasses major AI laboratories including OpenAI, Google, xAI, Anthropic, and Meta.
Frontier Models Qualify for Immunity Under Bill
According to OpenAI's statement on the matter, the company views this approach as focusing on meaningful risk reduction while enabling widespread access to AI technology across Illinois businesses and communities. Company representatives emphasized that such measures help establish uniform national standards rather than fragmented state-by-state regulations.
Unified Standards Over Fragmented State Regulations
The legislation carves out specific scenarios where liability protections apply, including instances where bad actors weaponize AI systems to create chemical, biological, radiological, or nuclear weapons. If an AI model independently conducts actions that would constitute criminal offenses and result in critical harms, developers would face no liability provided they published required safety, security, and transparency reports.
Carve-Outs for Criminal Weaponization Remain
OpenAI's Global Affairs team member Caitlin Niedermeyer testified in favor of SB 3444, advocating for consistent federal AI regulation. Niedermeyer highlighted concerns about regulatory fragmentation across states, arguing that patchwork requirements could create operational challenges without substantially improving safety outcomes.
The proposed legislation addresses a significant gap in the American legal landscape. Currently, neither federal nor state lawmakers have established clear liability frameworks for AI developers regarding harms caused by their technology. As artificial intelligence systems grow increasingly sophisticated and capable, these legal questions continue to gain urgency and relevance for industry stakeholders and policymakers alike.