FEATURE not yet become standard across the market. Many professional liability policies remain silent regarding AI.
However, carriers are beginning to pay much closer attention to the exposure. Rather than simply excluding AI altogether, we are starting to see insurers address the exposure through policy definitions, underwriting questions, coverage restrictions and, in some cases, sublimits.
I expect this area to continue evolving quickly, particularly when it comes to generative AI such as ChatGPT, Claude and Microsoft Copilot.
The distinction is important. Generative AI doesn ' t simply process existing information; it can create new content. That content might include reports, analyses, recommendations, legal research, policy comparisons, correspondence and other professional work products.
The professional liability concern arises when a professional relies on AI-generated content without adequately verifying its accuracy and then provides that information to a client.
WHEN AI GETS IT WRONG
Generative AI has been known to produce information that sounds authoritative but is inaccurate, or even completely fabricated. These errors are commonly referred to as AI " hallucinations."
We have already seen this exposure play out in the legal profession.
In the recent LNU v. Blanche matter before the Ninth Circuit, two attorneys filed briefs containing nonexistent cases, misattributed quotations and material misrepresentations of actual cases resulting from generative-AI hallucinations. Importantly, the issue wasn ' t simply that the attorneys used AI. The problem was that inaccurate AI-generated material was submitted to the court without being properly verified.
“
... AI may have created the error, but the resulting allegation against the professional could look very much like a traditional E & O claim.”
The consequences included monetary sanctions and a six-month suspension from practice before the court.
That distinction is extremely important from a professional liability standpoint.
AI may be the tool that creates the error, but the professional remains responsible for the product delivered to the client.
The same basic exposure can apply to virtually any professional. An engineer could rely on an incorrect building code generated by AI. A consultant could incorporate fabricated data into a client report. An insurance agent could rely on an AI-generated policy comparison that incorrectly states that a particular coverage is included.
In each case, AI may have created the error, but the resulting allegation against the professional could look very much like a traditional E & O claim.
AI COULD ALSO INCREASE CLAIM FREQUENCY
There is another side of the AI discussion that may ultimately be just as important to professional liability carriers.
SEPTEMBER / OCTOBER 2026
5