Artificial Intelligence and the Reconfiguration of Emotional Well-Being (2020–2025): A Critical Reflection

Between 2020 and 2025, rapid advances in artificial intelligence (AI) reshaped how individuals access emotional support, express feelings, and build interpersonal trust. This article offers a critical reflection—based on an analytical review of 40 peer-reviewed studies—on the psychosocial, ethical, and sociotechnical tensions that characterize AI-mediated emotional well-being. We document both opportunities (expanded access to support, personalization, and early detection) and risks (simulated empathy, affective dependence, algorithmic fatigue, and erosion of relational authenticity). Methodologically, we applied a three-phase critical review: exploratory reading, thematic clustering, and interpretive synthesis; sources were retrieved from Scopus, Web of Science and PsycINFO and filtered by relevance, methodological rigor, and topical fit. We propose a conceptual model integrating three interdependent levels—technological–structural, psychosocial–relational, and ethical–existential—and argue for a sociotechnical perspective that recognizes AI as a co-constitutive actor in emotional ecologies. The article closes with targeted research agendas and policy recommendations to foster human-centered AI that preserves emotional autonomy and equity.