Abstract
This systematic literature review examines the impact of artificial intelligence (AI) on academic integrity in higher education, synthesizing evidence from a curated set of 45 peer-reviewed records published between 2017 and 2024. As AI technologies - particularly large language models such as ChatGPT-become increasingly sophisticated and accessible, their implications for academic honesty, assessment authenticity, and educational integrity have escalated into a critical concern for institutions globally. The present review was not conducted as a conventional PRISMA flow‑based meta‑analysis with independently verifiable stage counts; rather, it is grounded in a carefully assembled source bank of 45 records identified through systematic searching, abstract screening, keyword filtering, and DOI verification. The analysis is organized around six major thematic areas: AI tools and academic dishonesty, detection technologies and effectiveness, institutional policy responses, student and faculty perceptions, ethical frameworks for AI in education, and future directions. Findings indicate that while AI offers meaningful pedagogical benefits, it simultaneously presents substantial challenges to academic integrity, with detection tools exhibiting limited reliability, notable biases against non‑native English writers, and an ongoing vulnerability to simple obfuscation techniques. Institutional policy responses remain fragmented and inconsistent, underscoring the urgent need for comprehensive, evidence‑based frameworks. The review further reveals that student and faculty perceptions are shaped by generational factors, AI literacy, and the clarity of institutional guidelines. By providing a detailed synthesis of current evidence, this review contributes to the growing body of knowledge on AI and academic integrity and identifies critical gaps requiring further investigation.

This publication has 40 references indexed in Scilit: