<![CDATA[This article examines the integration of artificial intelligence (AI) into the epistemology of science learning in primary schools through a literature review approach. The review focuses on how scientific knowledge is constructed, validated, represented, and negotiated in classrooms when AI is employed as a pedagogical support tool. Twenty-five national and international articles published between 2024 and 2026 were selected based on thematic relevance to primary education, science instruction, AI literacy, and epistemic dimensions of learning. The findings indicate that AI can strengthen conceptual exploration, learning personalization, scientific visualization, adaptive feedback, and students’ epistemic agency. At the same time, the review identifies risks related to cognitive dependency, informational bias, blurred knowledge authority, and weakened scientific verification when teachers do not guide AI use critically. The pedagogical implications highlight that AI should not replace scientific reasoning; instead, it should function as a mediator that supports students in asking questions, comparing evidence, testing explanations, and reflecting on how knowledge is produced. The review recommends strengthening teachers’ AI literacy, data literacy, and epistemic literacy so that AI integration in primary science learning remains ethical, reflective, and oriented toward the development of scientific reasoning from an early age. ]]>
Copyrights © 2026
