Does Insurance Cover AI? What Cyber and Liability Policies Say in 2026
No line of insurance fully covers AI yet, and most policies still say nothing about it. Here is our read of where AI risk sits in 2026.
AI losses have stopped being a thought experiment. A $1.5bn copyright settlement, a $243m Autopilot verdict and a $25.6m deepfake payment have all landed in the past two years. For brokers, underwriters, and claims handlers, the question is no longer whether AI creates insurable loss, but which policy pays when it does.
The honest answer in 2026 is that no line of insurance fully covers AI yet, and most policies, cyber included, still say nothing about AI at all. Cyber comes closest, because it usually responds to an attack whatever tool was used. On several liability lines, AI exclusions are starting to appear at renewal. Back in 2023 we called AI a new silent cyber risk. Three years on, the market has only started to answer, and the lines that have answered are answering differently.
Read the rest with a free account
The library is open to anyone with an account — articles, white papers, market updates and the masterclass archive. It is free, there is no time limit, and nothing renews.
Already have one? Sign in
More from Guides
What Has AI Already Cost? AI Losses Through an Insurance Lens
AI losses already run from a $31,000 court sanction to a $1.5bn settlement. The harder question for insurers: who paid?
When Cyber Insurance Starts Dictating Ransomware Response
Cyber insurance is often described as a financial backstop - a way to transfer risk once the worst has already happened. In…
SMEs Are Putting The Cyber Insurance Market Model Under Strain
SMEs, Pricing Discipline, and the Search for Sustainability