Chubb Is Excluding the Risk Its Own CEO Says AI Will Solve

Key Takeaways

• Verisk AI exclusions effective Jan 1 2026
• Major carriers filing AI exclusions with regulators
• GenAI lawsuits rose 137% YoY, 700+ cases
• Enterprise deployers now bear AI liability exposure
• MGA model better positioned for AI underwriting

Summary

Carriers are formally excluding generative AI liabilities from commercial general liability policies, with Verisk/ISO endorsements taking effect on January 1, 2026. At least six major insurers, including WR Berkley, AIG and Great American, have filed AI exclusion endorsements with state regulators. U.S. generative AI lawsuits surged 137% year‑over‑year in 2024‑2025, surpassing 700 filings and highlighting a rapidly expanding liability gap. The gap is being filled by specialist MGAs, which are better equipped than traditional carriers to underwrite AI‑related risk.

Pulse Analysis

The insurance industry is undergoing a structural pivot as standard carriers retreat from AI exposure. By adopting Verisk’s Form CG 40 47 and CG 35 08, insurers can strip generative‑AI‑related bodily injury, property damage, and advertising injury from both occurrence and claims‑made policies. This regulatory‑driven exclusion, already embraced by WR Berkley, AIG and others, signals that traditional commercial general liability (CGL) products are ill‑suited to price algorithmic failures. Consequently, a clear market segment has emerged for niche underwriters willing to model the nuanced risks of hallucinations, defamation and IP infringement.

Litigation data underscores the urgency of this shift. Testudo’s database records over 700 U.S. generative‑AI lawsuits between 2020 and 2025, a 137% jump in the most recent year. Claims cluster around patent and copyright infringement, as well as privacy‑related personal injury, with settlements like Anthropic’s roughly $3,000 per infringing work establishing early benchmarks. Compounding the problem, the four leading AI vendors—OpenAI, Anthropic, Google and Microsoft—limit liability to twelve months of fees, leaving the enterprise deployer to shoulder residual damages. This contractual architecture transforms a legal concern into an underwriting challenge.

Specialist Managing General Agents (MGAs) are poised to fill the coverage void. Their flexible capital structures and ability to craft bespoke policy language make them uniquely suited to price AI‑specific exposures, a task traditional carriers deem too volatile for their balance sheets. However, MGAs must secure reinsurance capacity and develop robust actuarial models; failure to do so could choke the nascent market. As AI adoption accelerates across sectors, the MGA model may become the primary conduit for AI liability protection, reshaping risk transfer dynamics for years to come.