2026 ASEE Annual Conference & Exposition

Generative artificial intelligence (AI) tools such as ChatGPT are increasingly used by engineering students as informal study partners, yet little is known about the quality of the instructional support they provide in upper-division aerospace courses. This paper examines how ChatGPT functioned as a study aid in unit in an undergraduate aeronautics course covering airfoil aerodynamics and thin airfoil theory. To approximate realistic student use, a scientifically literate but non-expert learner engaged in more than 175 exchanges with ChatGPT while studying instructor-developed course notes in preparation for a quiz. Explanations, diagrams, feedback, and practice problems generated by the AI were analyzed for technical accuracy, alignment with course learning goals, and quality of instructional support.
Four consistent patterns emerged: ChatGPT frequently (1) omitted or misrepresented key visual representations, (2) discouraged engagement with mathematical derivations, (3) provided overly affirming feedback that did not surface gaps in understanding, and (4) generated oversimplified practice materials lacking disciplinary precision. While responses were often conversational and reassuring, they emphasized efficiency over conceptual depth. For cumulative, math-intensive courses like aeronautics, this instructional stance may leave students feeling prepared without fully developing the underlying understanding needed for exams and subsequent coursework. The paper offers practical considerations for aerospace instructors navigating student use of generative AI tools.