Generative Artificial Intelligence and Academic Integrity in Higher Education: A Critical Policy Review

Abstract

The public availability of generative artificial intelligence tools since early 2023 has disrupted the assumptions on which academic integrity regimes in higher education were built, since those regimes were designed to detect the appropriation of existing text rather than the generation of novel text on demand. This article aims to review critically how the scholarly and policy response has developed, to evaluate the three dominant institutional strategies of detection, prohibition, and assessment redesign, and to assess the adequacy of the Indonesian regulatory framework for governing academic conduct under these conditions. The study is a critical, non-systematic review combining thematic synthesis of peer-reviewed literature published between 2023 and 2025 with document analysis of Indonesian regulations on plagiarism and higher education quality assurance; it does not report original screening statistics and makes no claim to exhaustive coverage. The review finds that the discourse has shifted from containment toward pedagogical redesign, that detection-based strategies rest on an evidential basis too weak to sustain disciplinary decisions, that blanket prohibition produces unenforceable rules and inequitable outcomes, and that assessment redesign is the only strategy whose effectiveness does not depend on distinguishing human from machine text. It also finds that the Indonesian framework, constructed around plagiarism understood as the appropriation of another person's work, does not straightforwardly reach undisclosed machine generation. Institutions are advised to regulate disclosure rather than use, to design assessment for process evidence, and to treat artificial intelligence literacy as a graduate competence.