The objective of this prospective, parallel-group, 1:1 randomized controlled trial is to evaluate the effectiveness of a generative artificial intelligence (AI)-based immersive training program, utilizing the Gemini platform, in improving endodontic risk disclosure and informed consent (RD-IC) quality, communication competence, self-efficacy, and reducing communication anxiety among undergraduate dental students during their clinical clerkship.

Traditional pedagogical methods in dental education, such as case-based learning (CBL) and peer role-playing, often struggle to simulate the high-stakes interpersonal tension and technical complexity involved in specialized endodontic risk disclosure. Complications like instrument separation, pulp floor perforation, and acute flare-ups require clear, empathetic, and legally sound communication. While standardized patients (SPs) are effective, they pose severe logistical and financial constraints for large student cohorts.

To address this gap, 96 fifth-year undergraduate dental students undergoing clinical rotation were recruited and randomly allocated to either the experimental group (1-month generative AI-based immersive training via the Gemini platform) or the control group (1-month conventional case-based learning). The AI intervention provided interactive, real-time conversational simulations across six high-stakes endodontic clinical scenarios, delivering instantaneous, personalized feedback on technical accuracy and empathetic delivery.

The primary outcome evaluated was the RD-IC completeness score. Secondary outcomes included SEGUE communication framework scores, Self-Efficacy in Clinical Communication Scale (SCS) scores, Communication Anxiety Scale (CAS) scores, blinded standardized patient (SP) ratings, user satisfaction, and long-term clinical skill retention assessed at 1-month (T2-1m) and 2-month (T2-2m) post-intervention follow-ups.