USC rethinks engineering coursework and assessment for AI

USC’s Ming Hsieh Department of Electrical and Computer Engineering is embedding AI into teaching while changing how students demonstrate their knowledge. The effort gained urgency after Stephen Tu showed colleagues AI-generated mathematical proofs that they could not distinguish from student work. Faculty are emphasizing critical thinking, foundational engineering concepts and the ability to identify errors and validate AI outputs.

Assignments now incorporate code generation and prompt engineering, including work in Bhaskar Krishnamachari’s courses that uses AI models to generate functional Python applications. Faculty stress that engineers remain responsible for design decisions and their ethical and human consequences. Feedback from industry partners has reinforced the importance of AI literacy in engineering workflows.

Assessment is shifting toward observable skills through in-person labs, physical measurements and interactive problem-solving. Mark Redekopp uses oral quizzes and hand-coding exercises with verbal explanations. Krishnamachari has introduced recorded video evaluations, using AI analysis to identify shared knowledge gaps and adjust subsequent lectures.

Rehan Kapadia and Yasser Khan have redesigned EE 105 around hands-on, project-based learning, much of it centered on AI. A custom knowledge graph helps students build the foundations needed for experimentation. Across the department, faculty are treating AI as part of the learning process rather than an outside shortcut, while retaining an emphasis on analytical rigor and human judgment.