Insurers hesitate at claims faced by OpenAI, Anthropic in AI lawsuits: report (seekingalpha.com)

OpenAI and Anthropic are weighing the use of investor money to settle potential multibillion-dollar lawsuits as insurers grow reluctant to offer broad liability coverage for AI-related risks, the Financial Times reports. While both firms carry traditional business insurance, underwriters are nervous about insuring exposures unique to large AI models — from copyright and defamation claims to privacy breaches, safety failures and systemic harms — because these risks are novel, potentially industry-wide, and lack historical loss data that insurers use to price policies. The hesitation matters for AI/ML because it could shift financial risk back onto companies and their investors, slow deployments, raise costs for startups, and spur demand for bespoke AI insurance and stricter internal risk controls. Technically, underwriters are re-evaluating how existing lines (E&O, cyber, D&O) apply to opaque, constantly updated models and may impose exclusions, higher premiums, or require demonstrable governance: audits, red-teaming, model documentation, and incident response. The gap highlights a broader need for clearer regulatory standards, standardized risk metrics, and actuarial frameworks for AI liability — developments that will shape how fast and safely advanced models are commercialized.