Insurance coverage for AI-related losses

Insurance coverage for AI-related losses can involve conventional business policies and products designed for AI performance risk. A facial-recognition coverage dispute, Citizens v. Wynndalco, concerned a business liability policy; Munich Re separately describes aiSure products for AI providers and deployers. The starting point is the insured loss and the wording of the relevant contract.

Mapping a loss to insurance

The following is a reading guide to potentially relevant coverage, rather than a determination that any particular AI claim is insured. California's insurance regulator advises reviewing coverage, exclusions and endorsements together when assessing business insurance.

• Bodily injury, property damage and advertising injury: commercial general liability policies address specified liability exposures. Their coverage grants and exclusions determine which allegations fall within the policy. AI involvement is one part of that factual analysis.
• Errors in technology products or professional work: errors-and-omissions coverage is a relevant starting point for a customer's alleged financial loss from a deficient technology product or service. Travelers describes technology E&O in terms of errors, omissions and negligent acts in the insured's products or services, and emphasizes matching the insured services to the business.
• Cyber incidents and resulting liability: the FTC distinguishes first-party response costs and business interruption from third-party claims, settlements and litigation costs. It recommends checking vendor-held data, geographical scope and defense wording.
• Media, computer fraud and social engineering: Travelers' June 2020 CyberRisk Tech sample separates media liability from computer-fraud, funds-transfer-fraud and social-engineering grants. These are distinct wording questions when evaluating, for example, a hypothetical defamatory AI output or impersonation fraud.
• Employment claims: The Hartford describes employment-practices liability coverage for claims such as discrimination, harassment and wrongful termination. Allegations concerning an AI-assisted hiring decision would require analysis of the particular insured act and exclusions.
• Management liability: review the applicable directors-and-officers coverage and endorsements where a claim concerns management's AI decisions or disclosures. The Berkley form discussed in AI exclusions and insurance policy wording expressly amends a directors-and-officers form, alongside employment-practices and fiduciary forms.

Defense, indemnity and performance

A duty to defend concerns supplying a legal defense; indemnity concerns payment of covered liability. In Wynndalco, applying Illinois law, the Seventh Circuit assessed potential coverage of the underlying allegations and affirmed a duty to defend. Its June 15, 2023 opinion provides a concrete example of why the defense issue deserves separate analysis.

AI-specific performance products address a further type of risk. Munich Re describes aiSure-backed performance warranties for providers and coverage options for deployers' AI-related financial losses. These descriptions identify a product approach; an issued contract supplies the insured performance measures, limits and other operative terms.

Related articles

• Anat Lior
• Insurance and AI-governance policy
• Insurance
• AI exclusions and insurance policy wording
• Insurance coverage disputes involving AI
• AI in insurance claims and underwriting disputes