Skip to content

US insurers add generative AI exclusions as regulators approve new forms

US insurers add generative AI exclusions as regulators approve new forms

Insurance filings across several US states show carriers beginning to weave in exclusions and policy updates tied to artificial intelligence, reflecting both ISO form changes and the escalating use of generative AI.

It’s one of those slow-burn shifts that suddenly becomes visible once regulators start posting the paperwork.

In Idaho, National Union Fire Insurance Company of Pittsburgh, Pa. introduced new endorsements for its hospice, home health care and related organisations program.

The updates span war, human trafficking and generative AI. Definitions now include an entry for generative AI, and the forms exclude cover for bodily injury, property damage and personal or advertising injury tied to the tech’s use.

AIG, National Union’s parent, said the filing relied on an ISO-standard form. The company also said it has no plans to implement the exclusions, suggesting the language arrived by default rather than intent.

National Union filed a similar package in Illinois. Regulators there pushed back, asking what real-world scenario the generative AI exclusion was meant to capture.

The carrier responded that AI applications stretch from basic generative tools to robotic workers and that claim potential will only rise.

It positioned the endorsement as a future underwriting lever, not an immediate shift in appetite. The filing also noted that Illinois had already approved an ISO form with near-identical language.

Berkley Insurance made comparable filings in Connecticut for private company management liability and crime.

Great American updated its Washington commercial umbrella and excess forms to exclude liability stemming from generative AI.

According to our analysts, we’re seeing the early outlines of a trend: insurers building optional AI exclusions now so they can pivot when loss patterns start to form.

Regulators approved all filings reviewed, which suggests they see the exclusions as consistent with ISO updates rather than a sudden contraction in cover. For buyers, the message is that AI-related risk is entering the underwriting vocabulary, even if the market hasn’t decided how aggressively to use these tools yet.