Academic integrity among higher education students in the era of generative artificial intelligence: pedagogical strategies and institutional approaches

Abstract

Purpose. To determine how higher education students use generative artificial intelligence, examine the associated risks to academic integrity, and assess the extent to which institutional and pedagogical response mechanisms align with student practices. Methods. The study employed a secondary analysis of published empirical data from 2024–2026. The study compared quantitative indicators across years, specifically identifying changes in percentage points. The study analyzed regulations governing the use of artificial intelligence in higher education institutions thematically. To summarize the findings, the study compared issues related to the use of generative artificial intelligence with institutional and pedagogical response mechanisms. The samples from different studies were not statistically combined. Results. British students' use of generative artificial intelligence in assessed assignments has risen to nearly 100 percent, and the direct inclusion of generated text is steadily increasing. In the U.S., the estimated prevalence of academic dishonesty is higher among daily users of generative AI and varies by field, although a causal relationship has not been established. It was found that the widespread use of AI is accompanied by a lack of AI literacy and insufficient support from instructors, and Ukrainian students (using medical education as an example), while noting the risks of unreliable information, remain undecided regarding the acceptability of using AI. The study systematized pedagogical strategies (developing AI literacy, critically evaluating content, establishing rules for use, and adapting assessment methods) and analyzed documents from Ukrainian higher education institutions. Rather than banning artificial intelligence, these documents call for disclosure, distinguishing practices in syllabi and human verification of individual contributions; however, they differ in how formally they are framed. The recommendations of Ukraine's Ministry of Education and Science and the Ministry of Digital Transformation set out principles for the responsible use of artificial intelligence. At the same time, the law on academic integrity requires disclosure of the use of generated results. We compared the problems and response mechanisms and found that most problems have a direct institutional counterpart. Conclusions. In the context of using generative artificial intelligence, ensuring academic integrity shifts from verifying the actual use of the technology to confirming authorship, which requires consistency in regulatory frameworks, competency development, and assessment design. Verifying a student's actual independence remains the least well-addressed issue.