Current trends and prospects of AI-assisted cognitive behavioral therapy: from the perspective of support and complementation

Cognitive behavioral therapy (CBT) has demonstrated efficacy for a wide range of mental disorders; however, access to treatment remains limited due to a shortage of trained professionals and institutional constraints. This article traces the development of digital CBT (dCBT) against this backdrop and examines recent advances in CBT support utilizing generative artificial intelligence (AI) and large language models (LLMs). Three primary applications are discussed: chatbots, personalization, and recommendation systems, with representative research findings presented for each. AI-powered CBT shows promise as an early intervention tool within stepped care models. Nevertheless, the therapeutic alliance—characterized by the capacity to share uncertainty—remains a uniquely human function that AI cannot replicate. Moving forward, it will be essential to distinguish between outcome indicators common to both human-delivered and digital CBT, such as symptom reduction and adherence, and those specific to human therapists, such as therapeutic alliance. This perspective may inform future research and clinical implementation of AI-assisted mental health interventions.