Aipedia: Veterinary surgeons and practices
Vets can use AI services to interpret radiographs and ultrasound, but the vet who acts on the report remains responsible for the diagnosis and must stay within their competence. A missed fracture or mass is a common source of claims, and an AI report does not move that responsibility.
Part of: AI in veterinary practice
In short
AI imaging services read radiographs, and in some cases ultrasound images, and return a report or flagged areas, often within minutes. Under the RCVS Code, vets must keep within their own competence and refer responsibly (1.2) and provide appropriate and adequate care (1.3). The RCVS Spring 2026 advice says clinical decisions must not be wholly delegated to AI and output must be critically appraised. If a fracture, mass or effusion is missed, the claim usually lands on the treating vet. Whether the AI provider shares liability depends on the contract, not on the PI policy.
Practices upload radiographs to a service that applies AI, sometimes followed by a specialist review, and receive a report with findings and suggested differentials. Some digital radiography systems now include on-screen flags for features such as cardiac enlargement or pulmonary patterns. Ultrasound AI is less mature but growing.
For a general practitioner without a radiologist on site, this can raise the standard of interpretation. The danger is the opposite effect: treating a report as a final answer when the images, history and examination point elsewhere, or when the study quality was too poor for the tool to read.
Turnaround matters too. A fast AI report can tempt a busy practice to discharge or start treatment before anyone has looked at the full study, and the report may cover only the region or question it was asked about. Incidental findings elsewhere on the image are still the treating vet's concern.
The relevant duties are:
Illustrative scenario: a cat is radiographed for a cough. The AI report describes a bronchial pattern and the vet starts treatment. The vet does not review the whole study, and a small mass at the edge of the image, outside the area the tool focused on, is missed. Months later a different practice identifies it, and the owner alleges a delayed diagnosis.
Similar problems arise when positioning or exposure is poor, when the tool was not trained on that species or view, or when a report on a subtle fracture is accepted without a second view being taken.
Communication can fail as well. If the owner is told the images were "checked by AI and clear", that statement may later be quoted back in a complaint as a promise the practice cannot support.
A missed-diagnosis claim is usually brought against the practice or vet that treated the animal, and the question is whether a reasonably competent vet would have identified the finding. PI required by Code 3.4 is designed for that kind of allegation, subject to the policy terms. The AI provider is rarely a party to the claim. If the fault lies in the service, any recovery usually depends on the supplier contract; see AI contract clauses and indemnities. Where a human specialist reviews the AI report, the terms on who is responsible for that review matter too.
You can reduce the risk with these steps:
See also AI differential tools, AI lab analysers and the AI in veterinary practice overview.
Apex arranges PI for vets and veterinary practices. A broker reviews your cover and how the policy responds where an AI imaging report formed part of the diagnosis.
Use it as one input. The RCVS says decisions must not be wholly delegated to AI and output must be critically appraised. Review the images yourself, check the report fits the clinical picture and record your own interpretation.
Possibly, depending on the contract, but the owner's claim is usually brought against the treating vet or practice. Any recovery from the provider tends to depend on the supplier terms, which often limit liability.
Not automatically. Code 1.2 is about referring cases responsibly when they are outside your competence. A software report is not the same as referring to a colleague who takes clinical responsibility, so consider referral separately.
The RCVS asks vets to be open and honest about how AI processes client data. Explaining that a service helped interpret the images, and that you reviewed them yourself, is usually clearer than presenting the result as an AI verdict.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.