Insurers are starting to exclude AI from their directors' liability contracts

W. R. Berkley excludes any "use, deployment or development" of AI from its D&O and E&O policies, while Verisk/ISO offers an exclusion clause to forms that cover more than 80% of the U.S. market.

The insurance market is starting to distance itself from the risks associated with AI. W. R. Berkley has introduced an exclusion in its directors' liability, professional liability and fiduciary liability contracts for any use, deployment or development of artificial intelligence, real or alleged. The wording is so broad that it covers just about any company that touches these tools.

The movement goes beyond an isolated insurer. In January 2026, Verisk/ISO proposed an optional clause to exclude claims related to generative AI from bodily injury and property damage coverage. These standardized forms serve as the basis for more than 80% of property and casualty policies in the United States: their evolution therefore guides the entire market.

The numerical context remains modest but is growing rapidly — the AIIDB incident database lists 713 incidents, and about 250 AI-related lawsuits are ongoing in the United States, with misinformation and deepfakes dominating the incident categories. A report by the RAND Corporation recommends that state insurance regulators mandate an explicit notice of what is covered, so that companies do not discover the extent of the exclusion at the time of the loss.