Research Urges Caution on use of AI Dental Record Keeping
- Details
- Published: Thursday, 23 October 2025 09:46
- Written by Peter Ingle
- Hits: 545
A newly published study into AI speech tools in dentistry has determined that they could revolutionise dental record-keeping, but added that caution is needed.
The new study from King’s College London concluded that AI automatic speech recognition (ASR) tools could dramatically improve how clinicians record patient information, with benefits to patients and dental teams.
The appeal of such systems includes potentially saving time and reducing the current administrative burden. Against this, although transcriptional accuracy of these tools is high, there were problems with more technical language. As a result the study, published in the Journal of Dental Research, concluded that their reliability is not currently sufficient to support their use without operator review.
The researchers tested ten different ASR systems to see how well they could transcribe spoken orthodontic clinical records into written text. The best-performing system was an experimental pipeline combining OpenAI’s GPT-4o transcription with a large language model for error correction.
The potential of AI assisted record keeping is considerable given the amount of time spent by dental professionals typing up clinical notes. As this tends to happen during the consultation, traditional note keeping reduces face-to-face time with patients. This is a problem that has only got worse as the standard for record keeping has steadily risen, in response to changes in techniques, materials, and medico legal expectations.
There is great potential for reliable ASR tools to allow clinicians to dictate their clinical notes naturally, and so allow them to focus more on direct interaction with their patient. The most advanced systems were faster and more accurate than manual typing, with up to 60% time savings.
The AI-enhanced experimental pipeline (GPT4oTranscribeCorrected) had the lowest error rate, especially with technical dental terms.
Some commercial systems including Heidi Health digital scribe and GPT4oTranscribe speech-to-text application programming interface, performed well, however others including Dragon Anywhere, had high error rates and even introduced clinically significant mistakes.
The study found that background noise and accent had minimal impact on the best systems, making them suitable for real-world clinical settings.
While the technology shows great promise, the research team warned that clinically significant errors such as misidentifying teeth on treatment plans, can still occur. As a result they recommend a “human-in-the-loop” approach, where clinicians review and edit transcripts rather than relying on them blindly.
Lead author Ruairi O’Kane said: “AI speech tools can streamline documentation and improve efficiency, but we must remain vigilant. Even subtle transcription errors can potentially impact patient care.”
Looking ahead, the team suggest that future systems should include confidence indicators to flag uncertain terms and be trained on larger, more diverse dental datasets. Ultimately, the goal is to help clinicians become editors of their notes, while maintaining safety and accuracy.
The study has been awarded the NIHR Integrated Academic Training Poster Prize across KHP at the inaugural KCATO Research Symposium, Best Clinical Presentation (DCT category) at the London, Kent, Surrey, Sussex NHS England Deanery event, and the British Orthodontic Society’s BOC Aspiring Orthodontist Poster Prize at the recent British Orthodontic Conference.
You need to be logged in to leave comments.
Report
My comments