Educational Purpose under AI's Technical Logic: Goal Substitution and Stage-Sensitive Governance in Basic and Higher Education
Artificial intelligence is becoming part of everyday teaching, assessment, and knowledge production in schools and universities. Debate has concentrated on the functions of these systems, but the educational consequences always depend on the logic through which they operate. AI applications are commonly organized around optimization, automation, prediction, datafication, and scale. Education follows a different set of commitments, including developmental appropriateness, sustained cognitive effort, professional judgement, human interaction, and responsibility for knowledge. Using critical conceptual analysis and directed content analysis of international policy documents and academic literature, this paper examines how these logics interact in basic and higher education. The analysis identifies goal displacement as the central problem. Indicators that are easy to measure and optimize can gradually redefine what institutions treat as learning. The consequences differ across educational stages. In basic education, premature cognitive outsourcing may interrupt the formation of foundational capabilities and reduce meaningful interaction. In higher education, AI-assisted knowledge production raises concerns about verification, authorship, disciplinary judgement, and academic responsibility. A Pedagogical Primacy Framework is proposed to guide educational decisions through four connected considerations: educational purpose, cognitive necessity, human agency, and accountability. The framework supports teacher-governed use in basic education and transparent, reviewable use in higher education.