Explore LLM-enabled Tools to Facilitate Imaginal Exposure Exercises for Social Anxiety

A team of researchers has published a paper exploring the use of large language models (LLMs) to support a specific Cognitive Behavioral Therapy (CBT) technique called Imaginal Exposure (IE) for social anxiety. Their tool, named 'ImaginalExpoBot,' was co-designed with mental health professionals to address a key challenge: traditional IE homework requires clients to construct and sustain their own detailed, anxiety-provoking narratives, which can be difficult. The AI tool generates these vivid, personalized exposure scripts automatically.

The research involved a formative evaluation with five therapists and a user study with 19 individuals experiencing social anxiety symptoms. The findings show that the LLM-enabled support can effectively help users prepare for real-world anxiety-inducing situations by providing immediate, user-specific scenario adaptation. Crucially, the AI-generated scenarios were reported to remain within a therapeutically beneficial 'window of tolerance,' avoiding overwhelming the user. However, participants and professionals identified limitations in the tool's continuity and depth of customization, highlighting the need for more adaptive, context-aware designs in future therapeutic AI applications.