Abstract:

Artificial intelligence (AI) has emerged as a powerful tool that can be applied to various activities, transforming the way we approach complex problems. Its application offers significant potential to optimize the design, production, and resolution of problems and tasks across different fields. This paper analyzes the possibility of using a generative AI (ChatGPT 3.5) in the field of university teaching, particularly in the field of engineering. The main objective of this study is to evaluate to what extent ChatGPT 3.5 can serve as a personalized tutor in undergraduate engineering courses. Specifically, we examined the ability of ChatGPT 3.5 to explain complex concepts related to the subject Signals and Systems usually taught in Telecommunications Engineering degrees. In addition, we investigated whether ChatGPT 3.5 can solve theoretical questions and typical problems related to this subject by providing learners with correct, adequate, and understandable explanations or justifications. We have selected a set of typical exercises and tested how well ChatGPT 3.5 solves them. These exercises involve fundamental concepts and common techniques used in the Signals and Systems course. When an inaccurate or even wrong solution was found, we varied the way these exercises are presented to ChatGPT 3.5 to obtain, if possible, a correct solution. In case ChatGPT 3.5 failed to obtain an acceptable solution, we explored whether this generative AI was able to indicate useful hints to solve the problem. In this way we obtained a set of rules or hints that could be useful for students to interact with ChatGPT 3.5. On the other hand, an experiment was conducted on the use of this generative AI by students enrolled in the current Signals and Systems course. Students were asked to solve a homework assignment by consulting ChatGPT 3.5. The students then self-assessed their homework. Subsequently, a survey was conducted on the use and help obtained by the AI. A first result was that 95% of the students admitted that they had previously used ChatGPT 3.5 for academic purposes. Another aspect to highlight is that, even though approximately only one third of the students considered that ChatGPT 3.5 helped them to solve the exercises, most of them (56%) were inclined to continue using this tool in their academic tasks. With this study, we aim to contribute to the integration of generative AI tools into the university academic environment, making them valuable resources for students.

Keywords:

Education, Artificial Intelligence, Undergraduate.