Students’ Ethical Reasoning about Generative AI in Higher Education Assessment: A Cross-National Survey Study of Romania and Moldova

Generative AI (genAI) has spread through higher education since late 2022, raising urgent questions about assessment practices and academic integrity. Yet students’ ethical reasoning about these tools remains under-examined. This cross-national survey study compares how Romanian and Moldovan students perceive and reason about genAI in academic assessment.

We collected data from 189 students (146 Romanian, 43 Moldovan) at two pilot institutions during 2024-2025. Students show complex reasoning: 95% use AI and 70% endorse it for education, but 58% worry about plagiarism alongside concerns about critical thinking, privacy, and fairness. Romanian students report higher AI familiarity (χ²=13.00, p=.011, V=.26) and stronger perceived benefits for research skills (Mann-Whitney U: M=3.35 vs. M=2.98, p=.042, r=.15) and content creation (M=3.32 vs. M=2.70, p=.002, r=.23). Benefits and concerns did not correlate (r<.07, p>.37), showing students can hold both positive and critical views simultaneously.

Both groups want clear institutional policies, AI literacy education, redesigned assessments, and ongoing dialogue about responsible use. These findings challenge both prohibition and unrestricted approaches, instead supporting context-sensitive policies that distinguish among AI types, assessment contexts, and teaching purposes.

This pilot study adds empirical evidence to debates about AI integration and offers practical guidance for educators and policymakers in contexts with different technological development levels.