CODING AND ARTIFICIAL INTELLIGENCE IN ELEMENTARY EDUCATION: A SYSTEMATIC REVIEW AND EVIDENCE MAP (2021–2026)

Abstract

Coding and artificial intelligence (AI) are increasingly introduced in primary education, yet evidence remains fragmented across computational thinking, AI literacy, digital literacy, ethics, and classroom implementation. This systematic literature review and evidence mapping synthesises primary-school studies published between 2021 and 18 July 2026. A verified screening log was developed through structured keyword searching, publisher and DOI verification, and backwards–forward citation checking. Eligibility was limited to empirical studies involving primary-school students, teachers, or schools and directly examining coding, AI learning, AI literacy, or classroom implementation relevant to pedagogical deep learning. Twenty-three studies were included. The evidence shows a shift from perception-oriented research toward concrete instructional designs, including block-based machine learning, unplugged–plugged sequences, collaborative inquiry, digital storytelling, AIoT, physical computing, analogy-based pedagogy, and debugging. Computational thinking and AI literacy were the most frequently reported outcomes. Stronger learning experiences shared three features: visible AI mechanisms, authentic problem contexts, and structured reflection. Evidence concerning long-term effects, equity, standardised assessment, and explicit deep-learning frameworks remains limited. The synthesis proposes a mechanism linking teacher readiness and infrastructure with pedagogical processes and student outcomes. These findings imply that primary-school implementation should prioritise teacher readiness, age-appropriate scaffolding, meaningful problem solving, and reflection rather than technology use alone, providing direction for curriculum design, teacher development, and educational policy.