Growing AI liability risks outpace policyholders' ability to address them
(The Insurer) - As artificial intelligence is increasingly deployed by companies and their employees, the risks of unintentional damage related to the technology rise, with insurers likely to carry some of the cost.
Incidents of harm from AI are increasing over time, according to the AI Incident Tracker, a project that categorises information from a crowdsourced database using a taxonomy of AI risks developed at the Massachusetts Institute of Technology.
While some of these are relatively minor, others involve physical harm to people, property damage or significant financial loss.
Among the more severe events in the database are incidences of alleged bias in algorithms used by job advertising platforms, an industrial robot fatally injuring a worker after mistaking him for an object and a deepfake scam tricking an executive to transfer millions to a fraudulent account.
Occurrence of AI-related litigation is also on the rise, according to insurtech Testudo. It reported an 81% year-on-year increase in generative AI-related lawsuits from January to April 2025. The median awarded damages for settled AI lawsuits were $4 million.
A Munich Re report last year found that AI-related lawsuits have gradually increased over the past 20 years, but recent technology trends caused increases in certain categories of litigation.
Increasing AI use in housing decisions was behind an increase in discrimination suits in 2018 and 2019, the report said, while self-driving vehicles triggered a rise in bodily injury lawsuits in 2020 and GenAI led to more recent jumps in IP infringement and defamation actions. "As the use cases for AI continue to proliferate through the economy, we expect the number of lawsuits to increase," Munich Re's report said.
While companies are more aware of possible impacts of AI since ChatGPT's launch in 2022, the risks are not yet well understood by policyholders.
Aon said business leaders were underrating AI risk when they ranked it 49th in a survey of current risks in 2023. It jumped to 17th place when the same participants were asked about their top future risks, suggesting companies saw it as more of a risk for the future than the present.
In Allianz's 2025 Risk Barometer, based on a survey conducted in late 2024, "new technologies" was ranked as the 10th major business risk, with 10% of those surveyed placing it among their top three. This was an increase from the previous year, but down from some past years such as 2019 when 19% saw technologies as one of their three biggest risks.
In the short term, consideration of AI risks appears to be crowded out by more familiar threats such as cyber incidents, natural catastrophe and disruption to supply chains.
This may be partly due to the challenge of knowing how best to manage AI risks. The Federation of European Risk Management Associations conducted a global survey in 2024 which found AI use was the risk organisations felt was most inadequately addressed, with 43% of respondents expressing concern.
A WTW survey in 2024 found respondents identified AI as the top emerging risk, a term which varies in scope but often refers to particularly unfamiliar and uncertain risks. Organisations "struggled to articulate how AI might change over the next two years," WTW said, with respondents defining the risk in dozens of different ways.
This week, The Insurer is publishing a series of articles about how the industry is beginning to grapple with AI's effect on liability risks.
Look out for articles every day on topics including the extent to which insurers are measuring their AI exposure, disputes over coverage, the effects of market cycles and whether AI risk could emerge as a product class in its own right, similarly to how cyber did in recent decades.
