Architects and engineers have spent two years asking whether AI tools belong in their workflow. Insurers have now answered a different question
Architects and engineers have spent two years asking whether AI tools belong in their workflow. Insurers have now answered a different question — whether AI losses belong in their policies — and for a growing number of carriers, the answer is no. The problem for design firms is that this change does not announce itself. It arrives quietly, as a form number on an endorsement schedule most policyholders never read.
Key Takeaways
Firms that want the endorsement-level breakdown can start with this guide to the Verisk AI exclusion and whether it is already on your policy, then use the checklist below at renewal.
Verisk, whose ISO forms supply the standardized policy language most U.S. property and casualty carriers build on, released two optional endorsements with a January 2026 edition date. CG 40 47 excludes bodily injury, property damage, and personal and advertising injury arising out of the use of generative artificial intelligence. CG 40 48 applies the same exclusion to personal and advertising injury only.
The definition is the part worth reading twice. Verisk defines generative AI as any machine-based learning system or model trained on data that can create content or responses—text, images, audio, video, or code. That language is broad enough to reach rendering tools, specification generators, and drafting assistants, not just chatbots.
Because the forms are optional, no carrier is required to adopt them. But standardized language removes the main barrier to adoption: carriers no longer have to draft and file their own wording state by state.
The Verisk forms attach to general liability policies, but professional liability carriers are not waiting. According to reports from Hunton Andrews Kurth and the National Law Review, Berkley has introduced an absolute AI exclusion for D&O, E&O, and fiduciary liability policies — one that specifically names ChatGPT, Bard, Midjourney, and DALL-E. Philadelphia Insurance and Hamilton Select have already excluded AI-related claims from certain E&O products. AIG and Great American have filed for regulatory approval to add comparable language.
"Absolute" is the word to watch. An absolute exclusion turns on any actual or alleged use of AI, which courts tend to read broadly. A claim with even a tangential connection to AI can fall within it.
For most firms, the honest answer today is probably not — yet. A policy written before 2026 generally contains no AI exclusion, and a claim involving AI-assisted work is treated as ordinary professional negligence. The picture changes at renewal. Here is the check, in order:
Firms already reviewing their professional liability insurance program should fold this check into that review rather than treating it as a separate exercise.
The liability logic has not changed: the licensed professional who stamps or seals the work bears responsibility for its accuracy, whether or not AI assisted. What changes with an exclusion is whether the E&O policy responds when that responsibility turns into a claim.
That gap matters most for firms whose use of AI is informal. Tools adopted team by team, without documentation or verification procedures, create exposure the firm cannot see — and, under an absolute exclusion, coverage the firm no longer has. Carriers have shown a willingness to negotiate carve-backs for supervised AI use when a firm can document a governance program that covers tool approval, verification steps, client disclosure, and licensed sign-off.
AI use also touches exposures outside E&O entirely — data handling, system compromise, media liability. Firms mapping their AI footprint should look at cyber liability insurance considerations in the same pass, since AI-related restrictions are appearing in that market too.
What is the Verisk AI exclusion?
A pair of standardized endorsement forms (CG 40 47 and CG 40 48, January 2026 edition) that allow carriers to exclude losses arising from generative artificial intelligence from liability policies.
Does it automatically apply to my E&O policy?
No. The forms are optional endorsements. They apply only if your carrier adds them — which is why the endorsement schedule, not the headline, is what to check.
Which carriers have added AI exclusions?
Berkley has introduced an absolute AI exclusion across professional and management liability lines; Philadelphia Insurance and Hamilton Select have excluded AI-related claims from certain E&O products; AIG and Great American have filed to follow.
Is AI-assisted design work still covered today?
On most in-force policies written before 2026, yes — claims are handled as standard professional negligence. The answer can change at renewal.
What should firms do before renewal?
Check the endorsement schedule and definitions, document AI tool use and verification procedures, raise the question with a broker in writing, and shop alternatives if an exclusion appears.
Not sure where your firm stands? Risk Specialty Group works with more than 20 "A" rated carriers specializing in architects, engineers, and design professionals. Request a comprehensive 360° risk review to map AI-related gaps across your full program before your renewal date makes the decision for you.
