Commercial insurance programs built before artificial intelligence entered workplace safety workflows may now contain coverage gaps. A white paper from Reed Smith and Haven Safety AI identifies where those gaps are most likely to surface in claims. Haven Safety AI co-authored the paper and sells an AI-assisted investigation platform. The insurance analysis draws on independently verifiable market developments.
ISO endorsement CG 40 47 01 26 became available for carrier attachment to commercial general liability policies in January. Filed by Verisk, the standard-forms arm of ISO, it removes coverage for bodily injury, property damage, and personal and advertising injury arising from generative AI. Two narrower variants also took effect: CG 40 48 removes advertising injury only, and CG 35 08 targets products and completed operations coverage.
Adoption moved quickly. W.R. Berkley, Chubb, Travelers, Berkshire Hathaway, and AIG had each filed to adopt the ISO forms or proprietary equivalents by April. A nationwide review of state filings found 2,369 generative AI exclusion records already in force across 49 states by mid-2026. The endorsements close the "silent AI" problem - policies that never expressly named AI, leaving coverage to be contested at claim time.
The Reed Smith paper frames the exposure around workplace safety. An organization using AI in incident investigations or corrective action programs could face a commercial general liability (CGL) claim where an AI-generated output contributed to a worker injury. Under CG 40 47, an insurer may deny coverage if the loss arose from generative AI. The trigger is the connection to AI, rather than the proportion of work it touched.
W.R. Berkley has introduced an absolute AI exclusion across its D&O, E&O, and fiduciary products. AI-specific E&O exclusions are appearing more broadly across the professional liability market. Safety professionals using AI-generated investigation content or corrective action recommendations should review policy language before a claim tests the wording.
Aon's 2026 AI risk report, produced independently of Haven Safety AI, adds the underwriting dimension. Carriers evaluating commercial accounts are increasingly looking for documented AI governance frameworks. Organizations that can demonstrate human oversight of AI outputs and formal approval protocols are better positioned on coverage and terms.
Mapping AI use to specific coverage lines before renewal is more likely to surface exclusion language while there is still room to negotiate. What neither the paper nor the market has answered is where replacement coverage sits once the ISO forms remove CGL protection. A nationwide review of state filings found no admitted standalone product available to replace the coverage the exclusions remove.
