Insurers expected to introduce GenAI liability exclusions: Evercore ISI

Evercore ISI expects more insurers to introduce policy language specifically excluding claims arising from generative AI (GenAI) to protect against silent coverage.

In a recent report, Evercore ISI noted that the growing use of GenAI and large language models (LLMs) has insurance coverage implications, as users and others are increasingly seeking damages for harm resulting from its use.

Liability insurance coverage for AI risk is attracting more attention given sizeable lawsuits against model providers for financial loss and personal injury due to faulty LLM outputs, copyright infringement, and discrimination/bias.

Insurers are also paying attention to Verisk’s 4Q24 emerging issues survey of insurers, which found that most respondents expect GenAI will result in product liability claims within the next one to two years (32% are moderately concerned, and 23% are extremely concerned).

As a result, insurers are introducing exclusions for claims arising from GenAI in policies to protect against silent coverage, and Verisk plans to introduce standard exclusionary policy language in an endorsement by 2026.

Evercore ISI said it is unaware of any insurance coverage for litigation involving GenAI model developers such as OpenAI or Anthropic, or businesses that use GenAI. However, the company believes these policyholders could argue for third-party coverage under tech E&O, product liability, GL, and EPLI policies, or first-party coverage under cyber and potentially property (business interruption). Evercore ISI sees tech E&O policies as most exposed to potential claims, although it considers the probability low at this point.

Evercore ISI said, “While it is early and there do not appear to be any insurance claims associated with negative LLM outcomes, it is an emerging risk that we are monitoring as AI adoption increases. Recent policy exclusions now being introduced by insurers/Verisk policy language shows insurers view potential liability as a possibility and are protecting themselves against silent cover. We expect more insurers will introduce policy language to specifically exclude claims arising out of genAI to protect against this risk, even though there have not been losses yet (unlike silent cyber coverage within GL/Property policies which initially resulted in losses before policy language changed).

“We suspect large account insurers like CB, AIG, Zurich, AXA XL and Allianz are most exposed to this novel potential risk but FT reporting suggests most potential losses are uninsured.”