In health sciences education, developing clinical reasoning and decision-making skills is essential. However, undergraduate students often face limited opportunities to engage in realistic, risk-free clinical practice before treating real patients.

The SIMULAI project (Interactive Medical Simulation Using Artificial Intelligence Language Models) addresses this educational gap by enhancing clinical competencies through virtual patient simulations powered by Adjustable Agents of Generative Pre-trained Transformer (GPT) models. The initiative involved fourth-year Dentistry students in designing and interacting with AI-driven virtual patients, simulating diagnostic interviews and treatment planning in a controlled digital environment. The activity was integrated into a clinical subject and facilitated by trained instructors.

Following the activity, a total of 41 students voluntarily completed a structured satisfaction survey using a 5-point Likert scale. Results showed high ratings in terms of interest (Mean = 4.46), clarity of expectations (Mean = 4.29), skill development (Mean = 4.15), usability (Mean = 4.54), and instructor preparedness (Mean = 4.44). Collaborative learning and peer interaction were also positively evaluated (Mean = 4.22). However, students identified a need for improved realism in virtual patient interactions (Mean = 4.02).

Psychometric analysis demonstrated high internal consistency (Cronbach’s alpha = 0.931). Factor analysis revealed two principal components:
(1) general satisfaction and attitude toward the activity, and
(2) perceived educational value and realism, jointly accounting for 65% of the total variance.
Cluster analysis identified distinct student profiles, reflecting varying levels of satisfaction and critical engagement.

Students suggested improvements in realism and conversational coherence of virtual patients, refining the pedagogical structure and timing of the simulations, and incorporating visual aids such as clinical images and radiographs. Ethical concerns were also raised regarding the integration of artificial intelligence, emphasizing the importance of using AI as a tool to support, not replace, clinical reasoning.

Overall, SIMULAI effectively promotes core clinical skills, collaborative learning, and digital competence, aligning with educational quality standards. The findings indicate student acceptance and highlight opportunities for the refinement of this innovative teaching approach.

Keywords: Simulation, Clinical, Patients, AI, Satisfaction.