2026-09-24 · Source: Risk & Insurance Magazine
Summary in 3 Points • RAND's report found that 84% of AI incidents involved misinformation or deepfakes since ChatGPT's launch in November 2022. • 60% of U.S. generative AI lawsuits focus on intellectual property disputes against AI developers rather than model output harms. • Insurers are increasingly excluding AI-related losses, with Verisk/ISO's exclusionary language appearing in over 80% of U.S. policies. --- Businesses are integrating artificial intelligence into various sectors, yet insurance coverage for related harms remains inconsistent. RAND's report highlights that 84% of 713 tracked generative AI incidents involved misinformation or deepfakes. In contrast, 60% of U.S. generative AI lawsuits focus on intellectual property disputes rather than harms from AI outputs. Insurers are responding differently, with some major carriers moving to exclude AI losses entirely, while others are expanding coverage to address specific AI-related issues. The report also notes that exclusionary language from Verisk/ISO is now present in over 80% of U.S. property and casualty policies. Meanwhile, companies like Munich RE and AXA XL are offering expanded policies to cover AI-related risks. RAND suggests that the lack of clarity in existing policies complicates the understanding of coverage for AI-related losses, urging regulators to standardise AI coverage notices and improve tracking of AI incidents. For the London Insurance Market, the growing focus on cyber liability insurance is crucial, as firms like Lloyd's of London adapt to emerging AI risks.
For the London Insurance Market, the growing focus on cyber liability insurance is crucial, as firms like Lloyd's of London adapt to emerging AI risks.