Governing Artificial Intelligence in Indonesian Higher Education: Academic Integrity and Regulatory Reconstruction

Abstract

The rapid diffusion of generative artificial intelligence into higher education has destabilised the assumptions underpinning academic assessment, while Indonesian regulation continues to conceptualise academic misconduct through categories developed for an earlier technological era. This study aims to examine how existing Indonesian legal instruments conceptualise academic integrity, to identify normative principles suitable for contextual adaptation, and to formulate the regulatory and pedagogical reconstruction required for defensible institutional governance. The research applies normative juridical method combining statutory, conceptual, and comparative approaches, analysing primary legal materials including the National Education System Law, the Higher Education Law, the Personal Data Protection Law, and ministerial regulations on plagiarism and quality assurance, supported by international normative instruments and peer- reviewed scholarship. The findings demonstrate that the prevailing definition of plagiarism presupposes an identifiable antecedent author and therefore does not reach machine-generated text, producing a norm that is formally intact but functionally hollow; that detection- centred enforcement generates foreseeable injustice and unexamined data protection exposure; and that the protected interest requires reconceptualisation as the correspondence between a credential and a competence. The article proposes a principle of graduated authorial disclosure, under which required disclosure is proportionate to the epistemic significance of assistance and impermissible use is defined by reference to the competence each assessment certifies. It recommends revision of the definitional architecture of integrity regulation, institutional adoption of tiered disclosure policies supported by data protection impact assessment, procedural guarantees against determination by detection score alone, and sustained investment in assessment redesign capability.