Barriers and Enablers for Generative Artificial Intelligence in Clinical Psychology: A Qualitative Study Based on the COM-B and Theoretical Domains Framework (TDF) models.

Abstract

Background: This study investigated the perceptions of care psychologists regarding the adoption of generative artificial intelligence (GenAI) in therapeutic practice. A qualitative method is used, relying on the Theoretical Domains Framework (TDF) and the COM-B model to analyse both perceived factors such as rejection (barriers) and factors that promote acceptance (facilitators) of GenAI. Methods: A thematic qualitative study design was adopted, and semistructured and in-depth interviews were conducted with 14 private care psychologists. The interviews focused on TDF domains to identify barriers and enablers. All the interviews were recorded and transcribed. The data were analysed via a thematic approach, with the identified topics associated with the TDF and COM-B components. Results: Twenty-three factors were identified that influenced the decision to accept or reject GenAI in therapy, with 14 factors acting as barriers and 9 acting as facilitators. These factors are classified within the TDF domains. Highlighted barriers included a lack of understanding of AI and concerns about the confidentiality and privacy of information shared in therapy, whereas the main facilitators were training in AI skills and the possibility of having a digital assistant. Conclusion: This study reveals the need for greater understanding of training in AI among psychologists. The acceptance of AI varies depending on the training and experience of professionals; some show concern for the future of their profession, whereas others highlight that it is an opportunity to improve interventions. Information privacy concerns are significant and have been identified as key factors for enabling AI deployment.