Ethical implications of artificial intelligence-based academic integrity systems in Palestinian higher education: a systematic review

Abstract

An examination of ethical concerns is necessary for analyzing AI's use in detecting and deterring academic dishonesty in higher education. Although AI-powered systems (incorporating natural language processing, stylometry, and machine learning classifiers) could offer solutions to academic integrity concerns that would scale, such systems are vulnerable to algorithmic bias against non-native English speakers, constrained transparency in algorithmic decision-making, and privacy and academic freedom violations from increased surveillance. A systematic literature review was undertaken of English-language peer-reviewed literature over the period 2012–2024, identified across six databases (Scopus, Web of Science, ERIC, IEEE Xplore, ACM Digital Library, and PhilPapers) and hand-searched reference lists, drawing on theories of AI ethics, procedural justice, epistemic injustice, and critical algorithm studies. Four major themes emerged from the literature review, which included fairness and algorithmic bias, transparency and accountability and due process, student privacy and surveillance and agency, and context-dependent justice in under-resourced settings. No study in the corpus was based in, or focused specifically on, Palestinian higher education; the implications drawn for Palestinian universities are therefore theoretical extrapolations from evidence generated elsewhere, not a synthesis of local empirical findings. All current ethical guidelines were intended for resource-rich western institutions and disregarded the specific context in which universities face resource and digital inequality. Six principles of governance of ethical AI are provided, and an emphasis is placed on how AI is a supplement to human judgment in academia and should therefore be implemented with attention to the specific context.