The policy names AI and says yes
Affirmative wording lists AI losses as covered, sometimes with a sublimit. Coalition added AI security events and deepfake funds-transfer fraud to its cyber trigger in March 2024.
Register checked 21 October 2026
The insurers, policies and exclusions that decide whether an AI mistake is covered. Every entry links to its source and shows the day we last checked it.
Every dated change in our log since March 2024, with the source behind each one. The log is more complete from 2025 onward. Exclusion filings run into the thousands at state level; the exclusions tracker counts them separately.
We stamp every policy, endorsement and exclusion with one of these. The wording decides the stamp, never the marketing.
Affirmative wording lists AI losses as covered, sometimes with a sublimit. Coalition added AI security events and deepfake funds-transfer fraud to its cyber trigger in March 2024.
Most older wordings say nothing, so cover is argued after the loss. It is why carriers now issue clarifying wording, like the affirmative AI wording Beazley added to cyber and tech E&O in September 2026.
An exclusion removes AI claims in part or in full. ISO forms CG 40 47, CG 40 48 and CG 35 08 let carriers exclude generative AI from general liability from January 2026.
Who sells AI cover right now, newest change first. A blank limit means the insurer has not published one, never that we guessed.
Insurer and form-bureau moves to write AI out of existing cover, dated and sourced.
Pick how your company uses AI and which policies you hold. See which exclusions to look for and the question to put to your broker.
Educational only. No quotes, no sales calls, nothing stored.
Check my coveragePlain answers to the questions buyers ask, each built on the register and cited line by line.
We do not sell insurance or take commissions, and no insurer pays to be listed. Nothing here is insurance or legal advice; talk to a licensed broker before you buy. Read the methodology.