2026 ASEE Annual Conference & Exposition

In a data science classroom, certain skills are immutable: reading a technical journal article, critically analyzing text or code not written by oneself, communicating in the discipline, and metacognition. Additional skills are becoming increasingly relevant with the proliferation of large language model (LLM)-based tools, including problem framing, prompt engineering, and systems thinking. This practice report presents three classroom activities that instructors may use as-is or adapt to design AI-centered tasks targeting these skills. Each activity is presented in a lesson-plan format with learning objectives, setup, implementation details, and assessment criteria. These activities were piloted across three graduate-level engineering and data science courses over two semesters. Observations reported here are informal and reflective in nature; a formal empirical study with IRB approval is planned for a future semester. Our goal is to reduce instructor preparation time and provide a replicable template for further instructional design.