ChatGPT and peers show 78% accuracy for traumatic dental injury guidance

A study in the Journal of the American Dental Association evaluated large language models (ChatGPT, Gemini, DeepSeek, Qwen) for answering traumatic dental injury queries. Across 200 clinical scenarios, the models achieved an average accuracy of 78%, with consistent adherence to current guidelines but occasional omissions. The results highlight the potential and limits of AI tools for rapid decision‑making in emergency dental care.