Abstract:

The growing presence of Artificial Intelligence (AI) tools in higher education is transforming how students search for information, obtain explanations, and solve problems. While these systems can generate structured responses quickly, their outputs may also contain inaccuracies, ambiguities, or fabricated information. Consequently, one of the current educational challenges is ensuring that students develop the ability to critically evaluate AI-generated responses rather than simply accepting them. In this work, we present the design and implementation of an educational experience aimed at promoting critical thinking and conceptual understanding through the evaluation of AI-generated answers in an undergraduate electronics course.

The activity was implemented in a second-year electronics subject within a Bachelor’s Degree in Physics. For many students, it represents their first contact with electronics concepts such as circuit theory, measurement techniques, and electronic instrumentation. These topics require both conceptual reasoning and analytical problem solving, making them suitable for exploring the potential and limitations of AI-generated explanations. Accordingly, the experience was designed with two main objectives: reinforcing students’ understanding of key electronics concepts and promoting a reflective and responsible use of AI tools in academic contexts. The activity follows the Challenge-Based Learning (CBL) methodology, which encourages collaborative work to address a defined challenge through inquiry, analysis, and discussion. In this case, the challenge consisted of evaluating the validity, clarity, and usefulness of answers produced by an AI tool in response to electronics-related questions. Rather than using AI simply as a source of information, students were required to analyze, question, and refine the responses generated by the system.

To support this process, the instructors prepared a set of questions covering different types of learning tasks. These included conceptual questions assessing the understanding of theoretical principles, calculation or problem-solving questions requiring several analytical steps, and methodological or design questions related to procedures or circuit analysis. This variety allowed students to explore how AI systems perform when dealing with different forms of reasoning in electronics.

The activity was carried out during classroom sessions in which students worked in small groups. Each group used an AI tool to generate answers to the proposed questions, which were then analyzed collaboratively. Students evaluated whether the explanations were conceptually correct, mathematically consistent, and sufficiently precise, critically examining vague explanations, incorrect reasoning steps, inconsistencies in calculations, or misleading interpretations. A key element of the activity was the iterative refinement of prompts. Students reformulated their questions and modified the contextual information provided to the AI system to observe how these changes influenced the quality of the responses. The results were compared across groups and discussed collectively with the instructors.

Overall, the experience suggests that integrating artificial intelligence within a challenge-based learning framework can promote critical thinking, strengthen conceptual understanding, and encourage a more responsible and reflective use of AI in higher education.

Keywords:

Active learning, artificial intelligence, challenge-based learning, critical thinking, electronics education.