A week of AI guidance lifted dental students’ practical scores

Dental students using ChatGPT during seven days of simulated practice scored 9.86 points higher on a practical test than an instructor-guided group. The randomised trial involved 123 students at one Chinese university and was published on 1 October.[1]

Both groups had unlimited opportunities to request guidance. Students used identical desktop VR equipment; the AI ran in a separate browser. Research assistants monitored its answers. Two experts, blinded to group assignment, assessed a subsequent 30-minute practical test.[1]

The outcome covered basic tooth preparation and immediate performance. Longer-term retention, patient outcomes and staffing savings were not measured.[1]

Why the comparison matters

There are two distinct questions an education experiment can ask. One is whether additional assistance helps someone learn. The other is whether a particular kind of assistance works better than an available alternative. A result is easier to use when the teaching decision it addresses is clear.

An earlier randomised trial by an overlapping Wuhan University team illustrates the first question. Published in May 2025, it involved 187 dental students and compared training videos plus ChatGPT with videos alone. Average practical scores were 73.12 in the ChatGPT group and 65.54 in the control group after a week. That was an encouraging result for adding conversational support, but the comparison did not pit the chatbot against access to professional instructors.[2]

The newer design therefore addresses a more demanding teaching choice. Its result supports the value of the supervised training arrangement.[1]

A school considering such a course should treat the arrangement as the unit to reproduce. An account subscription, a practice environment and a monitored learning programme are different purchases. The useful decision is how they fit together: what students practise, how questions are answered, and who resolves an explanation that seems inconsistent with the taught procedure.

Keep the outcome tied to the skill

For the next course, an educator could specify the assessment before choosing the assistant: the same practical task, marked consistently, followed by another assessment after a delay. That would make persistence of learning visible. A later stage could examine performance in a supervised clinical setting. These are proposals for building on a positive training result, each aimed at a different educational question.

The appealing possibility is assistance that makes practice more productive while leaving the learner responsible for the performance. A question can be answered on screen; the skill still has to be developed through doing. Evaluating that distinction is how a dental school can turn an attractive demonstration into a teaching method it understands.

Report an error or suggest a correction