The Impact of AI-Based Therapeutic Chatbots on Reducing Social Anxiety Symptoms: A Randomized Controlled Trial Comparing Artificial Intelligence Interventions to Traditional Cognitive Behavioral Therapy (CBT)

Social anxiety disorder (SAD), affecting ۷% of the global population, faces treatment barriers such as cost, stigma, and therapist shortages. This randomized controlled trial (RCT) compared the efficacy of AI-based chatbots with traditional cognitive behavioral therapy (CBT) in reducing SAD symptoms, examining longitudinal outcomes, domain-specific effects, and treatment adherence. A ۱۲-week double-blind study involving ۲۴۰ participants (aged ۱۸–۵۵) assessed outcomes using the Liebowitz Social Anxiety Scale (LSAS), biometric data, and adherence rates, analyzed via mixed-effects models. Results showed that AI-based treatment was non-inferior to CBT (effect sizes: AI d = ۱.۲ vs. CBT d = ۱.۳; p = ۰.۳۴), with comparable ۶-month maintenance (AI: ۶۸% vs. CBT: ۷۱%). AI excelled in reducing avoidance behaviors (p = ۰.۰۲), while CBT was more effective for cognitive restructuring (p = ۰.۰۱). Younger participants responded better to AI (۴۲.۳% vs. ۳۵.۱% symptom reduction; p = ۰.۰۳), and AI demonstrated higher adherence (۸۴% vs. ۷۲%; p = ۰.۰۲). Although therapeutic alliance was weaker with AI (p < ۰.۰۰۱), it did not affect outcomes. These findings position AI chatbots as a scalable, effective alternative for SAD, particularly for behavioral symptoms and accessibility, while CBT retains advantages in cognitive domains. A hybrid approach combining AI's reach with therapist-guided strategies may optimize treatment. The study informs stepped-care models and highlights the need for future research on cultural adaptations and long-term efficacy.

کلیدواژه ها:

AI-powered mental health interventions ، Cognitive Behavioral Therapy (CBT) efficacy ، Social Anxiety Disorder (SAD) treatment ، Therapeutic chatbots ، Randomized Controlled Trial (RCT)