Key points are not available for this paper at this time.
INTRODUCTION: Effective communication of health messages is essential. Pharmacy educators face the challenge of adapting curricula or integrating Generative AI (GenAI) tools. This study presents an educator-guided GenAI prototype, Authentic Teaching a control group). RESULTS: For students who had two data points for the primary outcome analysis, ATLAS users (188) showed significant improvement in their communication scores (median increase = 1.55 points, p = 0.0045), while controls (273) showed no change between years (p = 0.7). When comparing conversation attempts to OSCE scores in the ATLAS group (n = 201), improvements were seen with three or more attempts: three (+6.7), four (+6.5), five (+8.3), and six (+7.7). Domestic and female students benefited most. Students reported that GenAI avatars were realistic and enhanced learning, offering authentic, safe, and accessible practice with personalised feedback. DISCUSSION: Despite some technical limitations and AI hesitancy, ATLAS was a valuable supplement. The study supports that different communication subskills may be differentially associated with technology supported practice, and students who perhaps are not as confident to participate in the classroom could benefit from GenAI practice prior to classroom practice. However, many students still preferred in-person practice for developing empathy and nuanced communication. CONCLUSION: The present study provides associational evidence that educator guided practice with ATLAS correlates with higher OSCE communication scores, particularly for repeated users, amid notable student engagement and perceived value. Future research should prioritise multi-site randomised trials to minimise self-selection bias with explicit plans to assess equity impacts (e.g., for international students).
Lim et al. (Mon,) studied this question.