Assessing student knowledge in engineering courses has become increasingly challenging due to widespread access to external resources. Homework assignments and take-home assessments, which are often used to evaluate higher-level problem-solving skills, have long been vulnerable to solution manuals and online tutoring services. In upper-level undergraduate and graduate engineering courses, the complexity of problems frequently necessitates assessment formats that extend beyond standard in-class exam periods, further increasing reliance on take-home exams and projects.
More recently, the availability of artificial intelligence (AI)–based tools have introduced additional concerns regarding the reliability of take-home assessments. AI tools can generate complete and well-structured solutions to complex engineering problems, making it difficult to determine the extent to which submitted work reflects individual student understanding. This paper examines the impact of AI use on the assessment of student learning in undergraduate and graduate mechanical engineering courses.
Examples of assigned problems and corresponding AI-generated solutions are presented to illustrate cases in which AI produces correct, partially correct, or incorrect results. In addition, results from an anonymous student survey are analyzed to examine how students are using AI tools. Survey findings indicate that students increasingly view AI as a tutoring resource to support learning, while also acknowledging its use as a shortcut to obtaining problem solutions, which may limit learning. These results provide insight into both the opportunities and challenges AI presents for engineering assessment.
http://orcid.org/0000-0001-9829-2723
The University of Texas at San Antonio
[biography]
http://orcid.org/0000-0002-0811-9482
The University of Texas at San Antonio
[biography]
Are you a researcher? Would you like to cite this paper? Visit the ASEE document repository at peer.asee.org for more tools and easy citations.