Exploration of an artificial intelligence-supported training mechanism for teaching competence development in graduate students of prosthodontics

BACKGROUND: This study aimed to evaluate the effectiveness of an artificial intelligence-supported training framework in enhancing the teaching competence of prosthodontics graduate students. METHODS: Twenty-one first-year prosthodontics graduate students at Fujian Medical University participated in a structured, four-stage intervention over two semesters: baseline assessment, AI platform training, AI-assisted iterative teaching practice, and final evaluation. DeepSeek served as the core AI tool for lesson preparation, lecture script drafting, concept map generation, and interactive teaching design. Teaching competence was assessed at seven time points across five dimensions: teaching attitude and language, teaching content, logical structure, professional accuracy, and interactivity. Longitudinal changes were evaluated using repeated measures analysis, and students also rated the usefulness of the AI tool on a 5-point Likert scale. RESULTS: = 0.69). All five dimensions showed significant gains, with the largest improvements in teaching content and logical structure. Individual growth curves demonstrated continuous progress for nearly all participants. Students reported high perceived usefulness of AI (median 5, IQR 4-5), reflecting positive engagement with AI-assisted training. CONCLUSION: The AI-assisted training framework was associated with improvements in prosthodontics graduate students' teaching competence, particularly in content completeness and structural coherence. The introduction of DeepSeek, with strengths in Chinese semantic processing and professional reasoning, offers a potentially useful and localized tool for dental education. This framework may provide a strategy to help cultivate graduate students with skills in research, clinical practice, and teaching, potentially supporting the development of future high-quality faculty.