The effect of AI-enabled virtual patient simulation on training outcomes in psychotherapy education

Preparing psychotherapy trainees for clinical complexity is a persistent educational challenge. Virtual patient (VP) simulations enabled by artificial intelligence (AI) may enhance training by offering practical, hands-on experience. This study examined their effectiveness in improving training outcomes and reducing insecurity in future psychotherapists. Psychology students and psychotherapy trainees (N=87) were randomly assigned to four online 2D-VP sessions with automated feedback or pre-recorded role-play videos. Pre- and post-assessments measured psychotherapeutic competence, self-efficacy, knowledge, and insecurity. Participants with VP training significantly improved competence and self-efficacy, but not knowledge, and reduced insecurity, though effects did not surpass video training. Clinical experience predicted higher competence and lower insecurity but not training-related gains. Reductions in insecurity appeared earlier for video than for VP training. Simulation-related factors had no effect on training outcomes. Design recommendations emphasize pre-briefings and enhancing VP authenticity to use VP simulations as a scalable, engaging complementary method in psychotherapy training.