CFC makes AI’s role explicit in healthcare and professional liability cover

CFC announced explicit AI coverage language on 27 October for its global eHealth product and US Professions policy, with equivalent endorsements available for Professions policies elsewhere. The insurer says the wording clarifies coverage when AI forms part of an insured business and contributes to a claim.[1]

The development concerns a practical stage of AI adoption: a company has incorporated a tool into its service, something has gone wrong, and a customer or patient seeks a remedy. Understanding what service was delivered and what loss is alleged remains central to that account.

Making the policy’s intention visible

In its explanation of affirmative cover, CFC distinguishes policies that say nothing expressly about AI, policies that exclude it, and policies that define its coverage. The purpose of affirmative language is to make the insurer’s intended response explicit when AI is involved. That describes a policy approach, rather than a prediction about any particular claim.[2]

For healthcare, the announcement names AI errors, bias, hallucinations, ambient scribes and clinical-trial support software among the exposures addressed. It also says CFC’s healthcare package products have been upgraded. The announcement describes the expansion but does not supply the complete contract terms or settlement data.[1]

The service still shapes the claim

CFC’s published illustrations show why the underlying activity matters. In one, a patient alleges that a clinician relied on an AI diagnostic output and made an incorrect diagnosis. CFC describes protection for covered proceedings arising from healthcare services. In another, a professional firm passes inaccurate AI-assisted research into client advice, prompting an allegation of financial loss and professional negligence.[3]

These are insurer-authored scenarios explaining how cover is intended to respond. They are useful for following a chain from a model output to a delivered service and then to an alleged loss. They do not establish how frequently such events occur or provide a record of paid claims.[3]

The structure predates this announcement. CFC’s 2025 UK eHealth brochure separately describes bodily injury arising from healthcare or technology activities, technology errors and omissions, and cyber and privacy cover. It also lists separate technology and contract-related sections. That earlier document illustrates how several kinds of exposure can sit within a digital-health package; it is not the new AI endorsement.[4]

For an illustrative transcription service, the same incorrect sentence could matter differently depending on its use. A draft that someone corrects, a wrong statement delivered to a client, and an error incorporated into a clinical record describe different events. Following the sentence through the service is more informative than labelling all three an “AI failure”.

The commercially significant question is how the policy describes AI’s involvement in the service and the resulting loss. Following that sequence gives the coverage discussion a concrete starting point.[2]

Report an error or suggest a correction