Abstract
This systematic review examines the role of artificial intelligence (AI) in the development of sustainable mental health interventions through a comprehensive analysis of literature published between 2020 and 2025. In accordance with the PRISMA guidelines, 62 studies were selected from 1652 initially identified records across four major databases. The results revealed four dimensions critical for sustainability: ethical considerations (privacy, informed consent, bias, and human oversight), personalization approaches (federated learning and AI-enhanced therapeutic interventions), risk mitigation strategies (data security, algorithmic bias, and clinical efficacy), and implementation challenges (technical infrastructure, cultural adaptation, and resource allocation). The findings demonstrate that long-term sustainability depends on ethics-driven approaches, resource-efficient techniques such as federated learning, culturally adaptive systems, and appropriate human-AI integration. The study concludes that sustainable mental health AI requires addressing both technical efficacy and ethical integrity while ensuring equitable access across diverse contexts. Future research should focus on longitudinal studies examining the long-term effectiveness and cultural adaptability of AI interventions in resource-limited settings.