International Consensus on Artificial Intelligence in Undergraduate Dental Student Assessment: A Two-Round Modified Delphi and Live Consensus Study

Background/Objectives: Generative artificial intelligence (AI) creates opportunities for dental education but challenges the validity, authenticity and governance of student assessment. This study developed international consensus on principles and assessment-specific recommendations for AI use in undergraduate dental assessment. Methods: Two international electronic Delphi rounds were followed by a separately sampled live consensus exercise at the 2026 Association for Dental Education in Europe Annual Meeting. Policy consensus required ≥70% rating 7–10; assessment classifications required ≥70% selecting the same permission category. Free-text comments underwent qualitative content analysis. Results: Round 1 included 49 participants and 34/37 policy statements reached consensus. Round 2 retained 41 participants; all 19 revised/new policy statements reached consensus and 9/19 assessment formats reached categorical consensus. Ten unresolved formats proceeded to live voting (up to n=89); only virtual reality/haptics reached consensus (51/71, 71.8%, permitted with restrictions/guidance). Comments highlighted overlapping disclosure/restriction categories, intended learning outcomes, authenticity, enforceability, equity and patient-data safeguards. Conclusions: Stakeholders agreed strongly on governance principles but not on universal permission labels for complex assessments. An outcome- and risk-based framework with explicit rules, human accountability, protected evidence of competence, data safeguards and continuing AI literacy is supported.