Mitigating students’ reliance on generative AI in computer engineering education: a case study

The rapid evolution of artificial intelligence (AI) has a significant impact on teaching and learning, particularly with large language models (LLMs). LLMs were developed in the electronics and computer engineering sector, complicating the delivery of related courses such as coding and circuit design, as teaching AI is part of the lessons. While LLMs offer learning opportunities, they may lead to overreliance on automation, undermining traditional teaching methods. To address this, we implemented mitigation strategies in the curriculum to reduce dependence on LLMs, based on the “attack-defence” principle from game theory and hands-on problem-solving. Action research methodology is adopted to support iterative curriculum development and real-world experimentation in this study. These strategies were tested in a practical course. A three-phase survey, involving pre-course, on-course, and post-course assessments, was conducted to investigate student engagement and learning experiences. Survey results from 32 participants, with a full response rate in all three phases, indicate that learning experiences improved by 90.6% compared to pre-LLM scenarios, while mitigation strategies led to 84.4% better learning outcomes than unregulated LLM usage.