Major insurance carriers have already begun excluding AI from the policies boards depend on. WR Berkley filed an absolute AI exclusion across its D&O, E&O, and fiduciary liability forms in 2024, eliminating coverage for claims tied to AI use, including a director's own public statements about AI governance. Verisk, AIG, and Great American have since filed generative AI exclusions for general liability and umbrella policies, with new ISO forms that took effect January 1, 2026. None of these attach automatically, so a board's coverage may or may not have already changed depending on its renewal date. This piece explains what to confirm with your broker.
For years, AI risk in the boardroom has been framed as an adoption question: are we moving fast enough, are we falling behind. A quieter question has been settling into insurance filings at the same time, and it has nothing to do with regulation or an accreditor. It is whether your D&O policy will actually respond when an AI-related claim arrives.
What carriers have already changed
WR Berkley filed Form PC 51380 in June 2024, an absolute AI exclusion applied to its Directors and Officers, Errors and Omissions, and Fiduciary Liability forms. It eliminates coverage for any claim based upon, arising out of, or attributable to the actual or alleged use, deployment, or development of AI, by any person or entity. That scope reaches further than a technology failure. It also excludes claims tied to a director's own public statements about the company's AI governance or capabilities, so a board member's representation about AI oversight that turns out to be inaccurate can trigger a D&O claim this form does not cover.
WR Berkley is not alone. Insurers across the industry have moved in the same direction, and the pattern has extended into general liability. ISO released three generative AI exclusion endorsements that took effect January 1, 2026, which carriers attach to commercial general liability policies at renewal. Great American and AIG filed their own generative AI exclusion requests in late 2025. None of these attach automatically. They arrive one renewal cycle at a time, so a policy that already renewed earlier this year may carry the exclusion while one renewing later still might not, which is exactly why most boards do not know their coverage has changed until a claim tests it.
Why this reaches every regulated board, not just technology companies
A board does not need to build AI products for this exposure to matter. Once an AI system informs a decision the board is accountable for, from a retention algorithm to a diagnostic tool to a vendor-risk monitoring tool, the D&O policy is the protection directors and officers are counting on if that decision is ever challenged. If the policy now excludes anything AI-related, the protection the board believes exists is not the protection actually in place, and the institution's real exposure is larger than its risk committee has priced.
What to confirm at the next renewal
Ask your broker directly whether your D&O, E&O, or fiduciary liability policy carries an AI exclusion, and ask to see the endorsement language rather than a summary. Confirm whether the exclusion is absolute, reaching any AI use or representation, or narrower, tied to specific AI-caused harms. Ask what it would take to negotiate an exception or a separate AI-specific rider, since some carriers will write one on request even where the standard form excludes it.
The pattern behind the exclusions
Insurers are not waiting for a body of AI-related claims to develop before pricing the risk. They are pricing the uncertainty itself, and moving the exposure onto the policyholder ahead of it. For a board, that turns a governance gap into a coverage gap in the same motion. Naming the Human Authority Line™ for your AI systems, and being able to show who is accountable and how oversight is documented, is also one of the clearest ways to show a carrier you already understand and manage the exposure it is pricing against you. See what an engagement produces