Abstract
Depression and anxiety affect more than half a billion people worldwide, yet access to mental health care remains severely limited due to stigma, cost, geography, and workforce shortages. Internet-based cognitive behavioral therapy (iCBT) and, more recently, artificial intelligence (AI)-driven chatbots have emerged to bridge this gap. AI chatbots deliver 24/7 support through natural language processing and machine learning, simulating human-like conversations grounded in cognitive behavioral therapy principles. Evidence suggests that chatbot use improves engagement and reduces attrition compared with iCBT alone, with randomized trials and recent meta-analyses demonstrating short-term reductions in depressive and anxiety symptoms. Chatbots such as Woebot, Wysa, and Tess integrate mood tracking, automated check-ins, and structured therapeutic activities that enhance self-efficacy and emotional regulation. Despite these benefits, limitations remain, including the inability to replicate genuine empathy, risk of misinterpretation during crises, reliance on short-term evidence, and lack of standardized evaluation frameworks. Privacy and data security also represent significant ethical concerns. Future research should prioritize long-term, diverse studies, transparent reporting of therapeutic principles, and development of universal guidelines for safety and implementation. While AI chatbots cannot replace professional care, they represent an innovative and cost-effective complement to overburdened systems, particularly in resource-limited settings.