[stock_market_widget type=”card” template=”basic2″ assets=”VRSK” realtime=”true” api=”yahoo-finance”]
A single line of policy language can determine whether a business is protected or exposed when something goes wrong. That line is now shifting across the commercial insurance market, and most policyholders have not noticed. An endorsement filed by Verisk Analytics, Inc. (NASDAQ: VRSK), the standard setting organization known as ISO, has quietly attached itself to thousands of commercial general liability policies since January 2026, and it does one thing very clearly. It removes coverage for claims connected to generative artificial intelligence.
The endorsement, labeled CG 40 47 01 26, strips coverage for bodily injury, property damage, and personal and advertising injury when generative AI plays a role in the loss. Two related variants narrow that scope further, one focused only on advertising injury and another aimed at products and completed operations coverage. A nationwide review of state insurance filings found 2,369 of these exclusion records already active across 49 states, a pace of adoption that outstrips most prior coverage changes in the commercial lines market.
The carriers behind that pace are some of the largest names in the industry. W. R. Berkley Corporation (NYSE: WRB), Chubb (NYSE: CB), The Travelers Companies (NYSE: TRV), Berkshire Hathaway (NYSE: BRK.B), and American International Group (NYSE: AIG) had each adopted the ISO forms or built proprietary equivalents by April. That is a fast turnaround for an endorsement that did not exist a year earlier, and it signals how seriously underwriters are treating the risk that generative tools introduce into everyday business operations.
The exposure is not limited to companies building AI products. It reaches any business that uses generative tools somewhere in its workflow, whether that means drafting communications, running customer service chatbots, or supporting safety investigations. A white paper from the law firm Reed Smith, co-authored with Haven Safety AI, walks through a scenario in workplace safety. If an organization uses AI to help analyze an incident or recommend corrective action, and that output later contributes to a claim, an insurer holding the CG 40 47 exclusion may have grounds to deny coverage. The determining factor is whether AI touched the loss at all, not how central that role was.
The pressure is not confined to general liability either. W. R. Berkley has introduced a broad AI exclusion across its directors and officers, errors and omissions, and fiduciary liability products, and similar language is spreading through the wider professional liability market. Aon plc (NYSE: AON) published a 2026 report on AI risk that adds another layer to this picture. Underwriters are increasingly asking commercial accounts to document how they govern their AI use, and companies that can show human oversight and formal approval steps tend to get better terms.
What remains unresolved is what fills the gap these exclusions create. No standalone, state approved product currently exists to replace the coverage that CG 40 47 removes, which leaves a window between the risk companies are taking on and the protection their policies actually provide. Brokers who map out exactly where AI touches a client’s operations before renewal have a better chance of catching exclusion language while there is still room to negotiate or shop for alternatives.
The businesses most likely to benefit from this shift are the insurers and managing general agents willing to build products that address the gap directly, rather than simply excluding the risk and moving on. Those that design AI specific coverage or replacement endorsements early may find themselves with a meaningful edge as more commercial accounts discover, often only after a claim is filed, that the protection they assumed they had is no longer there.
