Decoding Podcast: Inside the Research on AI in STEM Education

Marielle Zhang, ‘25, hosts the podcast Decoding, exploring technology, learning, and the future of education. In this episode, she interviews recent Graduate School of Education (GSE) postdoc Karen D. Wang, Assistant Professor of Instructional Design and Technology at San José State University, about her research at the GSE on AI in STEM education.

Listen to the podcast

Runtime 43:13 | Read the transcript

Key takeaways

Study: Examining the Potential and Pitfalls of ChatGPT

Karen discusses her study examining the affordances and pitfalls of using ChatGPT in STEM courses. Her and her coauthors’ findings suggest that students can effectively use AI to identify relevant concepts needed to solve problems and explain those concepts in personalized ways to students. The study highlights that student learning should be focused on real-world, underspecified problems which are more difficult for AI tools to solve.

Read the article: Wang, K.D., Burkholder, E., Wieman, C., Salehi, S., & Haber, N. (2023). Examining the potential and pitfalls of ChatGPT in science and engineering problem-solving. Frontiers in Education, 8. https://doi.org/10.3389/feduc.2023.1330486

Study: Examining College Students Use of AI Tools

Karen shares another study that looks at how students are using AI in the real world. The study showed that students want to use AI to help them learn, but struggle to use AI tools in ways that fully support the learning process. Her and her coauthors’ found that around half the time, when looking for help on a problem, students input the problem into a chatbot and ask for a solution. This undermines students’ learning because the chatbot provides an answer right away. When using chatbots in this way, students don’t drive the decision-making process or have much insight into the process that leads to the proposed solution.

Read the article: Wang, K.D., Wu, Z., Tufts, L.N, Wieman, C., Salehi, S., & Haber, N. (2025). Scaffold or crutch? Examining college students' use and views of generative AI tools for STEM education. IEEE Global Engineering Education Conference, 1–10. https://doi.org/10.1109/EDUCON62633.2025.11016406 

Teaching Strategies

Marielle and Karen discuss strategies based on these findings that instructors can use to support students’ deep learning in the classroom. Here are some of the strategies Karen suggests:

• Discuss with students about how they can best use AI to support their learning, including modeling yourself how to use AI tools to constructively solve complex problems
• Share AI tools with students that are designed for supporting learning, instead of general-use chatbots (like ChatGPT)
• Articulate course learning goals and explain how students can use AI to support or undermine those learnings
• Spend time yourself using chatbots to investigate how AI could interact with your assignments
• Evaluate student performance with process-based rather than outcome-based assessments