Toward Human-Centered Artificial Intelligence for Users’ Digital Well-Being: Systematic Review, Synthesis, and Future Directions

Background As information and communication technologies and artificial intelligence (AI) become deeply integrated into daily life, the focus on users’ digital well-being has grown across academic and industrial fields. However, fragmented perspectives and approaches to digital well-being in AI-powered systems hinder a holistic understanding, leaving researchers and practitioners struggling to design truly human-centered AI systems. Objective This paper aims to address the fragmentation by synthesizing diverse perspectives and approaches to digital well-being through a systematic literature review. Using the stimulus-organism-response framework as a guiding lens, this study aims to develop a comprehensive model for designing human-centered AI systems that promote digital well-being. Methods A systematic review of 240 multidisciplinary publications was conducted to explore the intersection of AI, digital well-being, and human-centered design. The analysis involved identifying key themes, frameworks, and approaches, with the stimulus-organism-response model serving as an overarching perspective to organize findings and inform the model development. Results The review led to the development of a human-centered artificial intelligence model for digital well-being, a conceptual framework that consolidates current knowledge on designing AI systems to support digital well-being and positively influence human behavior. The proposed model integrates insights from cross-disciplinary research, providing a structured understanding of how AI system features (stimuli) affect users’ internal states such as perceptions and emotions (organisms) and lead to attitudinal or behavioral changes (responses). Additionally, this paper highlights emerging challenges and opportunities, including ethical considerations, scalability, and practical guidelines for applying the model in long-term research and practice. Conclusions This study contributes to advancing the field by presenting an overarching framework for fostering digital well-being through human-centered AI systems. By addressing gaps in the fragmented literature and proposing a unifying model, the findings offer insights for researchers and practitioners. The human-centered artificial intelligence for digital well-being model serves as a foundation for future exploration and practical application in creating intelligent computing systems that improve users' digital well-being in everyday life.