Anthropic filing flags AI losses that insurers may not cover

Anthropic’s IPO filing warned that some AI-related losses may be impossible to insure, underscoring a growing liability problem as companies deploy the technology more widely. In its S-1, the AI startup said model behavior that no user or developer intended could trigger substantial claims, while its existing policies may not be enough if one incident causes outsized harm. Anthropic also wrote that the market for AI-specific insurance is still young and that coverage for some liabilities may be unavailable or too expensive. CNBC reported insurers are already weighing broader exclusions as AI spreads through business systems and as claims tied to physical injury and even deaths draw closer scrutiny.