Generative AI in higher education: ethical and behavioral factors influencing students' intentions to use ChatGPT

Rizun, Bordean, Nikiforova, Beleiu, and Revina study a practical question for universities: what makes students intend to use ChatGPT for academic tasks, and how do ethical concerns fit into that decision? The paper is useful for AIEDHK because it does not treat generative AI adoption as a simple preference or novelty effect. Instead, it connects student intention with ethical trust factors, technology acceptance, planned behavior, and social influence. That gives educators a better vocabulary for discussing why students adopt AI tools and what kinds of support or safeguards may actually change behavior.

The study develops and tests an integrated model using survey data from 344 students in Estonia, Germany, Poland, and Romania. Its framework combines Fairness, Accountability, Transparency and Ethics in AI, the Technology Acceptance Model, the Theory of Planned Behavior, and the Unified Theory of Acceptance and Use of Technology. This combination matters because higher education use of ChatGPT is both technical and social. Students judge whether the tool is useful, whether they feel capable of using it, whether teachers and peers normalize its use, and whether they trust the system's privacy, explainability, and fairness.

One important finding is that ethical concerns shape trust, but trust itself does not directly drive students' intention to use ChatGPT. Explainability and privacy are reported as the strongest predictors of trust, while trust significantly affects perceived performance. In plain terms, students may care about whether ChatGPT feels understandable and privacy-respecting, but their decision to use it is more strongly shaped by whether they believe it will help them perform and whether they feel able and socially permitted to use it. That is a useful warning for institutions that rely only on abstract ethics statements.

The paper also emphasizes social influence and perceived behavioral control as key adoption drivers. University professors matter because their guidance helps define whether ChatGPT use is legitimate, expected, risky, or academically inappropriate. Perceived control matters because students who feel capable of using ChatGPT are more likely to adopt it. For AIEDHK, this suggests that responsible AI adoption cannot be handled only through bans or general policy notices. Students need concrete examples, boundaries, guided practice, and assessment rules that explain what good use looks like in a course.

The practical implication for Hong Kong higher education is that AI ethics and AI literacy should be designed together. Privacy, explainability, academic integrity, and fairness need to be taught through actual learning tasks, not isolated compliance messages. Teachers need course-level guidance so students know when AI use is permitted, when it must be disclosed, and how it should support learning rather than replace it. Product teams should also notice the model's message: trust is built through explainability and privacy, but adoption depends on performance, social norms, and student confidence. AIED systems should therefore make responsible use easy to understand, easy to practice, and visible enough for teachers to support.