FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Apex Insurance Brokers
Speak to a brokerGet a quote →

Aipedia: Dentists

AI radiograph analysis and caries detection

Dentists can use AI to flag caries, bone loss and periapical lesions on radiographs, but the recorded clinical evaluation must still be their own. Missed or overcalled pathology can lead to dental claims, and AI changes how those claims are argued.

In short

AI radiograph tools overlay suspected caries, bone levels and periapical areas on bitewings, periapicals and OPGs. Software that does this is likely to be regulated by the MHRA as a medical device. IR(ME)R 2017 regulation 12(9) still requires a clinical evaluation of each exposure to be recorded, and GDC standards 4.1 and 7.2 make the dentist responsible for accurate records and competent interpretation. GDC-commissioned research in October 2026 found AI least reliable in image-based tasks such as reading radiographs. Indemnity usually responds to the dentist's own negligent reading, but software defects may be excluded.

How dentists use AI on radiographs

Radiograph analysis is a widely discussed AI use in dentistry. The software reads bitewings, periapicals or OPGs and draws coloured marks over areas it classifies as caries, calculus, bone loss, periapical radiolucency or existing restorations. Some tools also measure bone levels for periodontal charting or produce a patient-facing image to explain findings. Practices may use them as a second reader, to speed up reporting and to help patients see what the dentist is describing.

The duties that apply

The GDC's October 2026 research on AI in dental education is directly relevant: AI was "least reliable" in image-based areas such as reading radiographs, and should support rather than replace diagnosis.

What can go wrong

Illustrative scenario: a dentist reviews routine bitewings with an AI overlay. The software marks two small enamel lesions but does not flag a distal lesion partly hidden by an overlapping contact. The dentist records "AI: 2 lesions, monitor" and moves on. A year later the missed tooth needs root canal treatment, and the patient alleges a negligent failure to diagnose.

The opposite also happens. The tool overcalls shadows or burn-out as caries, and fillings are placed that were not needed. Other failure points include poor-quality images the tool should not have been run on, findings outside the tool's scope (such as incidental pathology on an OPG), and an evaluation record that copies the AI output without any sign of independent reading.

How indemnity usually responds

A claim for a missed or overcalled lesion is usually handled as an ordinary clinical negligence claim against the dentist, and the question is whether a reasonably competent dentist would have spotted it. Relying on software is unlikely to be a defence on its own. Dental indemnity is compulsory under the Dentists Act 1984, and the GDC requires cover suited to your scope of practice. One defence organisation's published position says help with AI is available only where a human remains the decision maker, and it excludes software defects and product liability. A claim that the tool itself was faulty may therefore need pursuing against the supplier, which is uncertain. See does dental indemnity cover AI mistakes.

Practical controls

Our guide on what to document about AI use covers record-keeping in more depth. Related: AI periodontal and oral cancer screening, AI treatment planning and the dentistry AI hub. Doctors can compare AI in medical image reading.

How Apex helps

Apex arranges indemnity for dentists and dental practices. A broker reviews your cover and how it treats diagnostic decisions made with AI image analysis, including any software or product exclusions.

Sources

Frequently asked

Does an AI report satisfy IR(ME)R?

Not on its own. Regulation 12(9) requires a recorded clinical evaluation of each exposure. That evaluation should be the dentist's documented interpretation of the image. The AI output can be noted alongside it, but it does not replace the clinician's own reading.

If the AI missed it, is the supplier liable instead?

Possibly, but it is uncertain. The patient's claim is usually against the dentist, whose duty is to read the image competently. Any claim against the supplier depends on contract terms and product liability law, which one defence organisation says needs modernising for AI.

Should I tell patients that AI reads their X-rays?

There is no GDC rule requiring it yet, but valid consent under Principle 3 and the National Commission's recommendation on patient notification point towards being open. Practices can explain it when showing the image and mention it in the privacy notice.

Check your indemnity covers how you use AI

Tell us about your practice and the work you do. A broker will review how your current cover treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.

Start an indemnity quote Request a callback

Apex 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.