Abstract:

The introduction of generative artificial intelligence (GAI) tools into higher engineering education is gradually transforming the way students prepare for highly abstract theoretical courses. This change is particularly noticeable in disciplines such as Theoretical Electrical Engineering, where the need to combine conceptual understanding and practical application of complex mathematical models often creates significant difficulties.

The present study examines the impact of GAI on learning in the course "Theoretical Electrical Engineering", which covers fundamental aspects of the analysis of electrical circuits for direct and alternating current. The study is based on empirical data from an exam test conducted in 2025, including 630 student attempts. The large volume of data allows for precise analysis and identification of specific cognitive and conceptual gaps.

The analysis clearly defines “difficulty zones” where the success rate is below 40%. These zones include topics with a high degree of abstraction, such as the analysis of RLC circuits, working with phasors and complex impedances, three-phase power supply and resonance in electrical circuits. In contrast, success rates above 70% are reported in more basic and algorithmic topics, such as the analysis of direct current circuits and the direct application of Ohm’s and Kirchhoff’s laws.

On the identified problem topics, students are further prepared using GAI tools, applying various GAI use patterns.

Against this background, the study focuses on three questions, which are key for the development of future pedagogical strategies:
1. Is there a change in student performance in key difficult topics when artificial intelligence tools are used systematically as part of the learning process?
2. What GAI use patterns do students prefer in the learning process? The following GAI use patterns in engineering education have been considered: Calculator; Virtual Textbook; Step-by-Step Tutor; Simulation/Verification Tool; Case Study Generator; Jargon Interpreter; Soft Skills Development.
3. What are the pedagogical benefits (e.g. personalized learning, rapid feedback) and institutional risks (e.g. academic dishonesty, change in the role of the teacher) of integrating GAI?

The study applies a mixed methodology, combining quantitative analysis of test results with qualitative analysis of student practices.

Preliminary findings highlight the dual role of generative artificial intelligence. On the one hand, it has significant potential to aid understanding of the most challenging, abstract topics to build deeper knowledge. Students who use GAI show higher engagement and better results in difficult areas. On the other hand, serious risks include: generating inaccurate or wrong answers; promoting surface learning and circumventing the need for independent problem-solving, which undermines the development of critical thinking and problem-solving skills. On this basis, the article offers specific recommendations for pedagogical scenarios and institutional policies.

The final conclusion is that effective integration of GAI requires a rethinking of the curriculum and assessment, shifting the focus from the reproduction of solutions to the application and synthesis of knowledge.

Keywords:

Generative artificial intelligence, engineering education, theoretical electrical engineering, AI use patterns, AI-supported learning.