Abstract:

The irruption of Artificial Intelligences (AI) is going to have an impact on education, mainly in the way students search for information to study and do assignments. Traditional ways of searching for information, such as consulting books, journals, reliable internet sources or other bibliographic sources, will be affected by the ease of consulting a generative AI such as chatGPT. Although some research has concluded that these tools can answer short questions acceptably and even pass some university exams, the answers generated by AI may contain errors and may be inaccurate or unspecialised for the level required for university studies. Moreover, if consulting an AI replaces consulting lecturer notes and recommended books, teachers may lose control over the content that students study.

This article studies the students' perception of the validity and accuracy of chatGPT-generated answers to typical exam questions of the Manufacturing Technology course. After explaining the corresponding topics, a set of exam questions were proposed to the students from which they had to generate an answer using chatGPT 3.5. Subsequently, the students had to evaluate the validity of the AI-generated answers based on class notes, recommended books and other internet sources, generating a new answer and identifying the weaknesses of the initial chatGPT answer. The results are analysed together with the students' perception of the AI before and after the activity.

Based on the results, it can be concluded that this activity has helped students to understand the limitations of using generative AI to solve doubts related to Manufacturing Technology and has helped to improve their critical reasoning. It has also helped to reinforce the knowledge of the topics covered during the activity.

Keywords:

ChatGPT, AI, Artificial Intelligences, critical reasoning.