Shafran R, Bond L, Carlbring P, et al. From innovation to implementation: Artificial intelligence in cognitive behaviour therapy training and supervision. Behav Res Ther. 2026;197:104945. doi:10.1016/j.brat.2025.104945
This paper examines how artificial intelligence (AI) could improve the training and supervision of clinicians who deliver cognitive behavioral therapy (CBT). The authors explore AI’s potential benefits, the challenges of putting it into practice, and the safeguards needed to ensure it is used responsibly. There are several ways in which AI can support clinician training. For example, it could simulate patients through virtual conversations, allowing trainees to practice assessment and therapy skills in a safe environment where no real patients are involved. AI tools can provide immediate, personalized feedback, help trainees develop cultural awareness, and reduce the time and cost associated with supervision. Additionally, AI-based tools can also assist with case formulation by helping clinicians identify patterns across therapy sessions, recognize factors that may contribute to a person’s difficulties, and highlight information that could influence clinical decisions and supervision. While the benefits are clear, challenges exist in evaluating the clinical competence of AI. Determining competence requires clinicians to apply therapeutic skills flexibly to individual cases rather than simply follow treatment procedures. With this in mind, it is advisable that AI-assisted training be evaluated using observable measures of competence and tested across diverse clinical and cultural settings. Additional concerns about AI systems are also worth considering. AI systems rely on extensive collections of sensitive clinical data, presenting challenges related to privacy, data ownership, informed consent, and data quality. Successful implementation also depends on reliable digital infrastructure, financial resources, and implementation support, which may be limited in low-resource settings. Ongoing research, ethical oversight, and interdisciplinary collaboration are essential to ensure AI improves the quality, accessibility, and effectiveness of CBT training and supervision while minimizing potential risks.