Since the public release of ChatGPT in November 2022, the accelerated integration of generative AI (GenAI) into higher education has sparked widespread debate on how academic assessment should be redefined. While existing research extensively explores the technical capabilities of these tools, less attention has been paid to how students perceive and respond to the ethical challenges they raise.

This paper aims to address this gap through a dual approach: first, an empirical study of students’ perceptions regarding the ethical use of GenAI in academic assessment, based on a comparative analysis between 2 universities (higher education institutions) in Romania and the Republic of Moldova, and second, a set of practical recommendations designed to support students and educators in promoting academic integrity in a rapidly changing technological context.

The study is based on data collected via a questionnaire administered to a sample of 76 students. Findings reveal a strong correlation between students’ exposure to GenAI tools and their level of ethical awareness. Romanian students, who report more frequent use of such tools, show higher concern about risks such as misinformation, unintentional plagiarism, and the erosion of critical thinking. In contrast, students from the Republic of Moldova—still in early stages of adoption—emphasize the need for basic training and ethical guidance.

Grounded in both empirical data and a review of recent literature, the article proposes a framework of good practices relevant to both local and international contexts. For educators, the paper outlines pedagogical strategies for rethinking assessment, developing clear institutional policies, and embedding discussions on AI ethics directly into the curriculum. In doing so, this work contributes not only to a better understanding of how GenAI is reshaping higher education, but also offers concrete tools for a responsible transition toward AI-assisted learning and assessment.

Keywords: Academic integrity, ethical use of AI, higher education, AI-assisted assessment.