Preparing future dentists for artificial intelligence: a cross-sectional study of perceptions and educational needs in China.

Background: Artificial intelligence (AI) is entering dental education and clinical practice, yet little is known about how dental students at different stages of training understand its use, risks, and educational value. This study examined AI-related familiarity, attitudes, concerns, and learning needs among undergraduate, master's, and doctoral dental students in China. Methods: A single-center cross-sectional online survey was conducted in July 2025 at the School and Hospital of Stomatology, Wuhan University. The 20-item questionnaire collected information on demographic characteristics, AI-related exposure, and perceived application areas, attitudes toward AI, main concerns, responsibility attribution, and educational needs. Descriptive statistics were used to summarize responses, and Pearson's chi-square tests were used to compare differences across training stages. Results: A total of 343 valid responses were included, comprising 201 undergraduates, 92 master's students, and 50 doctoral students. Few respondents described themselves as unfamiliar with AI, but high familiarity with dental AI remained uncommon. Overall, 89.8% supported the use of AI in dental care, and 84.2% expressed interest in AI education. Training-stage differences were found in information sources, perceived application domains, suitable stages for AI use in clinical care, responsibility attribution, and expectations for future AI development. Undergraduates more often learned about AI through formal courses and tended to place AI near learning support and diagnostic assistance. Master's students showed greater acceptance of AI in treatment planning and implementation. Doctoral students were more cautious about AI in complex decision-making and gave more attention to responsibility, communication, and workload reduction. The most frequently reported concerns were data privacy and security, interpretability, and technological dependence. Conclusion: Dental students in this survey were generally open to AI, but their perceptions and educational needs differed across training stages. Dental AI education should move beyond general exposure and be aligned with students' progression through training. Undergraduate teaching may focus on basic AI literacy and early critical awareness, master's-level teaching on case-based evaluation within clinical workflows, and doctoral training on methodology, data governance, and ethical responsibility. Multicenter and longitudinal studies with practical assessments are needed to examine how AI-related competence develops during dental training.