The rapid adoption of generative AI, particularly ChatGPT, by young people can be seen as an evolution, or even a revolution, in educational practices. In higher education, our students, aged 18 and over, were born with the internet and new technologies. This Generation Z has specific characteristics, including a tenuous connection to high-tech. This has significantly changed the teacher-student relationship, as teachers are no longer the sole holders of knowledge. With a single click, students can access a huge amount of information and challenge their teachers on specific concepts.

While many studies explore issues related to the role of teachers (to be redefined), assessment methods (to be developed), and academic integrity (ethical issues, risk of plagiarism), our research stands out for its empirical approach focused on students. How do they perceive these changes? How do they use generative AI? Are they aware of the limitations of these tools? Our goal is to give students a voice so that they can express their perceptions and any fears they may have about generative AI, but also describe their actual and concrete uses of these new tools.

Our study is based on sixty interviews with higher education students. These students have distinctive characteristics: they are enrolled in different programs (business or management), at different levels (second or third year), and in either full-time or work-study programs. Data collection took place from April 2025 to June 2025. The results confirm that the majority of students surveyed say they are “always up to date” when it comes to technology in general, and only one student has not yet used ChatGPT. They use it in a professional context (for work-study students) as well as in a personal and academic context, several times a week or even several times a day. This diversity of practices reported by students reflects a pragmatic and reasoned approach to these tools. The lexical elements analyzed reflect an instrumental representation of generative AI that suggests a conception of AI as an assistant providing cognitive support, facilitating access to information and clarifying complex content. Students view the tool more as an aid than a substitute.

However, our results highlight an ambivalent attitude towards these tools. While positive aspects are clearly identified (time savings, assistance, speed, explanation of complex concepts), students perceive many negative aspects and have limited confidence in the results obtained. They highlight the loss of intellectual autonomy and excessively superficial answers. The issue of cheating is also raised, not to mention the fact that it encourages laziness. While demonstrating a high degree of lucidity towards these tools, students nevertheless find them very easy to use and do not feel they need training. In fact, their expectations express less of a need for institutional supervision than a desire for functional optimization of the tool.

The general attitude of students is one of reasoned confidence. The diversity of uses suggests a differentiated appropriation of the tool, mobilized not as an end in itself, but as a vector of efficiency, empowerment, and adaptation to academic, personal, and professional requirements. These tools undoubtedly represent an opportunity to be seized, provided that the objectives and rules of use are clearly identified by both teachers and students and that support is provided to make the best use of AI's potential.

Keywords: GenAI, uses, perceptions, higher education.