Generative Artificial Intelligence in Basic Education: A Critical Integrative Review of AI Literacy, Human-AI Pedagogy, and Curriculum Redesign
Year: 2026
Type: article
Abstract: Generative artificial intelligence (GenAI) has entered basic education faster than evidence, curriculum, and governance have developed. Existing discussions often combine generic artificial intelligence with GenAI, report immediate task performance as learning, and transfer findings from higher education to school-age learners. This critical integrative review examines how GenAI is being used in K-12 teaching and learning, what knowledge, skills, and attitudes it may support, which risks condition those outcomes, and what curriculum and teacher-development responses are warranted. The analytic corpus comprised 47 documents published mainly from 2020 to August 2026: 11 research syntheses, 16 empirical or contextual studies, nine policy and competency documents, and 11 foundational or methodological works. Sources were coded by educational level, AI role, learning activity, outcome domain, study design, risk, and implementation condition; empirical claims were weighted more heavily than conceptual or policy propositions. The evidence supports cautious conclusions. Structured, teacher-guided use can improve feedback, ideation, explanation, and short-term task performance, but durable learning remains uncertain and unguarded outsourcing may reduce independent performance. AI literacy therefore requires technical understanding, critical verification, ethical judgment, and learner agency rather than prompting skill alone. An author-constructed 5E application and a human-centered implementation framework position GenAI as a bounded tutor,collaborator, or assistant while preserving productive struggle and AI-free assessment checkpoints. For Thailand, adoption should be conditioned by Thai-language and cultural validation, child-data protection, rural infrastructure, and sustained teacher competence. The review concludes that curriculum redesign should optimize human-AI complementarity, not automate schooling.
Authors: Pia, Kaweechate, Kasatsutorn, Kittirat
Institution: Thaksin University
Language: Thai
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: diamond
OpenAlex ID: W7215628655
