Melanie Cooper

How Can We Use Generative AI to Support and Engage Students? 

Professor, Chemistry Department; 
Professor, Teacher Education and the CREATE for STEM Institute 
mmc@msu.edu 

A little more than two years ago, Generative AI (GenAI) chatbots such as Chat GPT emerged on the scene. Since that time, their use in education has skyrocketed. Surveys of students indicate that a majority use them to complete a wide range of assignments. However, there is emerging evidence that an indiscriminate reliance on such Generative AI tools may lead to reductions in critical thinking skills. Although this is not particularly surprising, it is important to acknowledge this issue and address it. As our understanding of the world becomes ever more complex, it is important that everyone understand how and why things happen, rather than simply knowing that they occur.  

“In the future, it will be ever more important that citizens are able to understand and use evidence to make arguments about how the world functions,” Cooper said. 

“Our work at Michigan State is focusing on the use of GenAI tools to support student scientific reasoning,” Cooper said. “We already know that students who engage in such activities are more likely to learn more deeply and are also more likely to be able to use their knowledge in new situations. However, such tasks are typically time consuming to grade or characterize, and particularly as the phenomena becomes more complex, it becomes difficult both for students to connect all the ideas needed, and for instructors to give them the support and feedback they need.” 

A set of generative AI chatbots—which can provide various types and levels of feedback to both students and instructors—are being designed and trained. These chatbots will be able to accurately characterize students’ responses and provide feedback in a variety of ways including 1) in the form of a Socratic dialog with students, 2) by scoring responses for both formative and summative assessment purposes, and 3) by providing summaries of student responses to drive instructional change.