Artificial intelligence (AI) applied to the educational field generates growing interest [1]. Many students use artificial intelligence in multiple tasks assigned to them. They are mostly looking for a reduction in the time dedicated to the task or looking for the information necessary to carry it out. It is also used in the resolution of problems and tasks that teachers request to be carried out as self-learning activities outside the classroom. Both pre-university and university students use these applications of artificial intelligence. The interaction with the latter through perception surveys on artificial intelligence shows their perception of AI is uncritical [2]. In order to train graduates with a critical vision of artificial intelligence, it is necessary to generate learning and/or evaluation activities in which students have to use AI. If the tasks are sufficiently complex, the subsequent interaction with the teaching staff allows the errors generated by the AI to be verified. Students, even university students, use freely accessible AI environments (programs, platforms) that are not created for relatively high-level educational environments. Therefore, on many occasions, they generate erroneous information and material. This work shows interactions carried out with students in which errors in the material generated by AI have been highlighted, errors of which, for the most part, the students were unaware. Problems and issues poorly resolved. Incorrect or incomplete references. On the other hand, it is noted that the text generally generated is not sufficiently specific to the learning field (Physics in engineering studies in the present study), being descriptive and generalist. Some examples will be shown in which these limitations are confirmed. Obviously, the evolution of AI is continuous and these errors are being reduced. On the other hand, AI specific to STEAM environments appears. However, some of the best software in development is not freely available. It should be noted that students, although they have experience in the application of AI, are users, and in most cases they are not even expert users.

References:
[1] H. Crompton, D. Burk. 2023, “Artificial intelligence in higher education: the state of the field,” International Journal of Educational Technology in Higher Education, vol. 20, pp. 1-22.
[2] J.J. Suñol, J. Daza, A. Wederni, L. Escoda.2024. “Artificial intelligence learning and assessment: perception of first-year university students,” EDULEARN24 Proceedings, pp. 5814-5819.

Keywords: Artificial intelligence, higher education, criticism, steam.