OpenAI Backs Bill That Would Limit Liability for AI-Enabled Mass Deaths or Financial Disasters (www.wired.com)

OpenAI has endorsed an Illinois bill, SB 3444, that aims to limit liability for AI developers when their models cause significant societal harm, such as mass deaths or financial disasters exceeding $1 billion. This marks a notable shift in OpenAI's legislative posture, as it moves from opposing liability-related bills to supporting one that could redefine industry standards. The bill proposes that developers of "frontier models," defined as AI systems trained with over $100 million in computational costs, can avoid liability for "critical harms" if they are not found to act intentionally or recklessly, provided they publish safety and transparency reports. This legislative push is significant for the AI/ML community as it could pave the way for similar measures nationwide, potentially influencing how frontier AI technologies are regulated. The bill responds to growing concerns over the safety of advanced AI systems, particularly as powerful models like Anthropic’s Claude Mythos are deployed. Although some experts express skepticism about the bill's chances, citing public opposition to liability exemptions and Illinois' history of stringent tech regulations, this effort highlights the urgent need for cohesive regulatory frameworks in the evolving landscape of AI technology.