Empowering mental health professionals in asynchronous online psychotherapy with GenAI

Background Asynchronous online psychotherapy (AOP) offers a promising solution to improve access to mental healthcare, particularly in times of global crises. However, therapists practicing AOP face challenges such as emotional strain, isolation, and increased workloads. The integration of generative artificial intelligence (GenAI) may provide support, yet little is known about its impact on the empowerment of mental health professionals, specifically regarding their perceived competence and autonomy, and how it can be integrated into blended care environments. Methods A mixed-method study was conducted with thirteen Ukrainian mental health professionals who engaged in both AI-supported and independent psychotherapy tasks using a digital agent, called APIA, powered by GPT-3.5/4. Participants completed standardized tasks such as forming diagnoses and treatment plans and responding to patient inquiries, followed by surveys and semi-structured interviews. Quantitative data were analyzed using descriptive statistics and paired-sample t-tests. Qualitative data from open-ended responses and interviews were analyzed thematically. Results Quantitative findings revealed that interaction with APIA led to a statistically significant increase in perceived competence (p < 0.01) and a marginal increase in perceived autonomy (p < 0.1). Qualitative data supported these findings, indicating that APIA enhanced task efficiency, supported validation of judgments, and improved independence. Three archetypes for integrating digital agents into AOP were identified: psychotherapist-centric, patient-centric, and therapy-centric. Despite the benefits, concerns were raised about data confidentiality, overreliance by inexperienced professionals, and potential depersonalization of care. Conclusions GenAI-based digital agents can positively influence mental health professionals’ sense of competence and, to a lesser extent, autonomy in asynchronous online psychotherapy settings. This empowerment may contribute to improved therapeutic quality and professional well-being. However, thoughtful integration, method-specific constraints, and ethical considerations must be addressed to realize the full potential of AI-supported mental healthcare.