,
Senior Lecturer, Tashkent State Medical University , Tashkent , Uzbekistan
,
Assistant, Tashkent State Medical University , Tashkent , Uzbekistan
,
Lecturer, Tashkent State University of Oriental Studies , Tashkent , Uzbekistan
,
Lecturer, Jizzakh State Pedagogical University , Jizzakh , Uzbekistan
,
Associate Professor, Termez State Pedagogical Institute , Termez , Uzbekistan
,
Lecturer, Department of Philology, Language Training Center, Alfraganus University , Tashkent , Uzbekistan
Lecturer, Kimyo International University in Tashkent , Tashkent , Uzbekistan
The conventional undergraduate psychology programs find it challenging to offer genuine clinical exposure because of the ethical considerations, the high prices of the standardized patients, and the lack of dynamism of the written case studies. The paper addresses the idea of integrating Generative AI (GenAI)-based clinical persona as a modification that can be scaled to close the gap between theoretical concepts and clinical environments. With the help of Large Language Models (LLMs) trained on symptom clusters and separate linguistic styles that correspond to the DSM-5-TR, a dynamic clinical sandbox was created and used to improve student skills. To measure two major metrics, Diagnostic Precision and Therapeutic Empathy, the study utilizes the mixed-methods approach. The diagnostic accuracy is measured with a Weighted Jaccard Index (WJI), which compares the student results with the programmed symptoms of a persona of the ground truth. Therapeutic empathy is evaluated by applying Natural Language Processing (NLP), whereby cosine similarity measures student dialogue to empathetic communication frameworks/theories that have been validated, e.g., the Person-Centered Theory developed by Rogers. Statistical results from a Randomized Control Trial demonstrate the efficacy of this framework, as students interacting with GenAI personas achieved a significantly higher diagnostic precision (Cohen's κ = 0.78, p < 0.001) vs. control (κ = 0.62) compared to those using traditional methods. A study was conducted randomized controlled trial (RCT) with 60 undergraduate psychology students comparing GenAI personas to static case studies. Furthermore, the experimental group demonstrated a 42% improvement in therapeutic empathy markers, confirming that interactive simulations facilitate superior skill acquisition. A more comprehensive ablation study also brings out the fact that empathy and realism scores go down considerably when algorithmic friction, as proposed by the Dynamic Resistance Logic, is not present. The paper wraps up with the ethical aspects, such as algorithmic bias and the uncanny valley of simulated suffering, which offers a roadmap to the future of AI-enhanced pedagogical systems in behavioral health education. Limitations include a single-site sample and text-based simulation; future work needs multimodal validation.
This is an open access article distributed under the Creative Commons Attribution Non-Commercial License (CC BY-NC) License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The statements, opinions and data contained in the journal are solely those of the individual authors and contributors and not of the publisher and the editor(s). We stay neutral with regard to jurisdictional claims in published maps and institutional affiliations.
