AI-powered CBT: Investigating the impact of voice-based generative AI on reframing anxious thoughts

With increasing demand for accessible mental health care, AI-driven interventions offer a scalable solution, yet their effectiveness in delivering cognitive-behavioral therapy (CBT) remains underexplored. This study examines whether AI voice-generated CBT exercises can help users reframe anxious thoughts. Participants engage with an AI-based CBT intervention, with pre- and post-intervention self-report measures assessing anxiety reduction and cognitive restructuring. Anticipated findings suggest AI interventions will be moderately effective, potentially comparable to human-guided CBT, though limitations in personalization and engagement may arise. These results could support AI’s role in expanding mental health accessibility, particularly for under-resourced populations. Poster submission was sponsored by Dr. Shahnaz Winer, (Psychology Department) and was presented at the Coquitlam campus on April 10, 2025, for Student Research Days 2025.