Truace
Subscribe

Artificial Intelligence as a Digital Consultant: Evaluating Large Language Models for Patient Inquiries on Dental Caries

ABSTRACT Objective To evaluate the clinical accuracy, usefulness, global overall quality (GQS), and readability of ChatGPT, Gemini, and DeepSeek responses to patient questions regarding dental caries. Materials and Methods Fifty‐nine simulated patient questions concerning caries diagnosis, etiology, treatment, and prevention were queried in Turkish. Three restorative dentists independently evaluated the responses for accuracy, usefulness, and GQS. Readability was assessed using the Ateşman index. Statistical ana…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-2189· cite & verify
Artificial Intelligence as a Digital Consultant: Evaluating Large Language Models for Patient Inquiries on Dental Caries
"09-8018-23" by NavyMedicine is marked with Public Domain Mark 1.0. To view the terms, visit https://creativecommons.org/publicdomain/mark/1.0/.

In brief

Researchers put 59 simulated patient questions about caries diagnosis, causes, treatment and prevention to three chatbots, ChatGPT, Gemini and DeepSeek, in Turkish. Three restorative dentists rated each response independently for clinical accuracy, usefulness and overall quality, and readability was scored with the Ateşman index.

The results suggest that chatbot choice matters when patients seek dental information, because the tools differed significantly in quality and in how easy their text was to read. The authors stress that these tools should complement professional advice, and the study does not show how patients would fare using them in real care.

Main points

  1. The test used 59 simulated questions on caries diagnosis, etiology, treatment and prevention, posed in Turkish.
  2. Three restorative dentists rated responses, and inter-rater reliability was high (ICC > 0.75).
  3. Performance differed significantly among the three models (p < 0.001), with Gemini scoring highest on accuracy, usefulness and quality.
  4. ChatGPT wrote the most readable text, generally rated easy to read on the Ateşman index.
  5. For DeepSeek, higher clinical accuracy correlated significantly with better readability.

The gain

Gemini produced the most accurate, useful and highest-quality answers to simulated patient caries questions, suggesting it could serve as a digital preconsultation aid.

The rundown

The study compared three widely available chatbots on the same set of caries questions. Dentist raters scored accuracy, usefulness and a global quality score, while a separate index measured how readable the answers were.

Statistical methods included the Friedman test for comparing models and Spearman correlation for relationships between variables. The models differed significantly, and the best performer on clinical quality was not the one with the most readable text.

The authors conclude that Gemini is the strongest candidate as a digital preconsultation tool among those tested. They add that AI tools should complement, not replace, professional dental advice.

Sources

  1. Peer-reviewedJournal of Esthetic and Restorative Dentistry2026-10-22

The debate