Who Is Liable When Artificial Intelligence Makes Mistakes?

Gretchen Hoff Varner’s commentary was featured in a Financial Times special report on how insurers and courts are grappling with liability for harms caused by artificial intelligence. The article explores a landmark case, involving allegations that an AI-powered hiring tool engaged in age discrimination, to illustrate broader questions about who bears legal responsibility when AI systems make mistakes. As companies increasingly deploy AI for critical business functions, the piece observes, insurers are seeking to exclude AI-related harms from standard corporate liability cover, much as they initially responded to past novel risks.

According to Gretchen, insurers’ approach to AI risks is following a familiar playbook from the early days of cyber threats. “This is an opportunity to sell more coverage,” she noted, explaining that as data breaches and other computer-enabled losses became more common, insurers made a business decision to carve those perils out of general liability policies and offer dedicated cyber insurance instead. She pointed out that a similar dynamic is now unfolding with AI, with insurers moving to restrict AI-related exposures under traditional policies, potentially paving the way for new standalone AI liability products.