Impact of AI-Generated Feedback on Dental Student Performance in Preclinical Prosthodontics Education

This study aimed to evaluate the impact of an artificial intelligence (AI)-generated formative feedback system on third-year dental students' crown preparation performance, perceptions, and burnout levels in a preclinical fixed prosthodontics course.

Fifty-eight students were eligible; two withdrew, leaving 56 randomized: AI Feedback Group (n = 28) and Faculty Feedback Group (n = 28). Five students were excluded from final analyses due to missing data, resulting in 51 students being analyzed. The AI system provided automated, multi-modal feedback on five critical parameters (occlusal/axial reduction, angulation, finish line quality, surface smoothness, and undercuts) via a 3D color-coded map, numerical score, and written comments. Academic performance was assessed via final practical exam scores; perceptions were collected via a 12-item custom survey; and burnout was measured using the validated School Burnout Inventory (SBI-9). Performance gains were compared based on post-hoc stratification by AI system usage frequency.

The mean final exam score for all participants was 84.3 ± 7.1(out of 100). A sub-analysis across the entire sample revealed that students with high AI engagement (≥ 5 uses; n = 24) scored significantly higher (87.2 ± 6.1) than those with lower AI engagement (< 5 uses; n = 27; 80.6 ± 7.2) (U = 388, p = 0.011). This highly engaged group also showed a greater mean reduction in flagged errors (27.8%) compared to the less engaged group (9.6%). Most students found AI feedback helpful (84%) and easy to understand (78%), with 76% reporting improved skill development. The preferred feedback method was a blended AI + faculty approach (69%). SBI-9 scores indicated moderate emotional strain (exhaustion = 3.6 ± 0.8) but low cynicism (cynicism = 2.0 ± 0.7); no significant differences in burnout were found between AI engagement groups.

Integration of AI-based formative feedback was positively received and correlated with significantly improved crown preparation performance in students who engaged more frequently with the system. AI-enhanced simulation supports self-regulated learning and autonomy in dental education without significantly increasing academic stress. This model offers a valuable complement to traditional instruction, supporting objective and timely skill development.