Educational Artificial Intelligence, Child Rights, and Human Care in Early Childhood

This article examines the use of artificial intelligence (AI) in early childhood education from a rights-based perspective, drawing on a critical interpretive synthesis (CIS) of the literature published between 2019 and 2025. A typology of four uses in early childhood —Tutor, Tool, Companion, and Tracker (THCR)— is proposed and each category is mapped against the core principles of the United Nations Convention on the Rights of the Child: privacy, non-discrimination, best interests, and participation. The contribution includes: (a) a risk-safeguard matrix differentiated by type of AI; (b) a logic model and theory of change for care-centered implementations; and (c) the SAFE LEARN checklist (Safety by design, Agency/assent, Fairness, Explainability, Learning alignment, Educator capacity, Accountability, Risk logging, Non-replacement of care). Implications for policy and practice are discussed for schools, administrations, and providers, emphasizing human mediation and verifiable equity as minimal conditions for acceptance. This work offers a pioneering framework that connects international normative principles with operational instruments, providing immediate guidance for the Education 2030 agenda and Sustainable Development Goals (SDGs) 4 and 16.