AI Misuse in Education and Academic Integrity Risks: Scoping Review with ☸️SAIMSARA

This scoping review aims to characterize the scope, determinants, and consequences of AI misuse in educational settings, and to synthesize the evidence on detection, prevention, and pedagogical countermeasures across diverse learner populations, disciplines, and institutional contexts. The review uses 127 references and builds its evidence map from 228 original studies with 144869 total participants/sample observations (topic-deduplicated ΣN). AI misuse in education appears to be less an isolated individual failing than a structural vulnerability, arising where highly capable tools, student pressure and rationalization, and weak institutional guidance converge. AI systems answered at least 65.8% of examination questions correctly across technical and natural science courses, highlighting how conventional unsupervised assessments can be completed without genuine learning. Detection tools showed accuracy ranging from 55.29% to 98% across tools, which suggests that detection alone cannot safeguard assessment validity and supports a role for assessment redesign, AI literacy, and explicit institutional policy. Because the evidence is dominated by cross-sectional, self-reported surveys, these associations should be interpreted cautiously rather than as causal effects. Future longitudinal and controlled studies should test whether redesigned assessments and clear AI-use policies reduce observed misuse while preserving students' critical thinking over time.