Earlier this fall, Dr. Minna Ng sat down with me to discuss her thoughts about generative AI and teaching. This is the first in a series of posts in which Learning Innovation and Lifetime Education will speak to instructors across the university to capture their insight into this emerging trend in teaching and learning. Thank you, Dr. Ng, for your time.
Minna Ng teaches and advises undergraduate students in the Neuroscience program at Duke. Below she speaks to the use of AI in the course NEUROSCI 102: Biological Bases of Behavior. She is co-teaching this Team-Based Learning course this semester with Dr. Nicole Schramm-Sapyta and they have two sections with 90 students each. Dr. Ng also reflects on past semesters of teaching and her general experiences with generative AI over the last two years.
How do you integrate generative AI into students’ assignments and your own work?
Professor Ng tends to use generative AI as a tool for teaching rather than incorporating it directly into her students’ assignments. One task she asks AI to perform is to help expand upon an idea for an assignment, such as breaking down the assignment into steps students can follow. While she may not adopt more than 10% – 15% of the AI-generated text outright, it quickly inspire more thoughts and productive text.
She also prompts AI to suggest analogies and stories for her lectures that reference current pop culture, making the content more relatable for students. For example, when preparing a lecture on the emotional brain, she asked AI to create analogies around the characters in the film “Inside Out,” who personify emotions.
She primarily uses Copilot for her AI explorations because she finds it easy to use and it automatically provides follow-up questions to refine the AI’s answers. For her needs, it didn’t seem worthwhile to compare many AI models.
Professor Ng teaches a team-based learning course, meaning students spend most of their class time working as a team to answer questions about the day’s content. This approach leaves little room for students to use generative AI. However, AI was incorporated into one assignment where students investigated neurological disorders. Students could use AI (with attribution) but were also taught how to find reputable sources, such as looking for .org or .edu URL extensions. This technique was meant to help students compare any content or sources they may have obtained from AI to accepted sources. In class, students shared research about disorders within their teams and had to decide as a group which one deserved a million-dollar research grant, minimizing reliance on AI.
How do you talk to students about AI?
For Professor Ng, it is important to have a blanket AI policy in the syllabus, but also to bring it up in other contexts throughout the semester. She noticed that on the first team feedback survey, students were using generic (possibly AI-generated) language to describe their teammates’ contributions. Professor Ng mentioned this trend in class and encouraged students to provide honest and thoughtful feedback, emphasizing that the course is a safe space for peer evaluation.
What advice do you have for instructors exploring AI and teaching?
Her advice for faculty is to stay informed through various sources, including articles or podcasts, to better understand AI and how to integrate it into teaching practices. She also suggested following experts in your field or colleagues who approach AI from both perspectives (for and against AI). Another starting point is to interact with generative AI with the goal of becoming proficient in prompting, which allows you to refine the output to make it increasingly useful for your work.
Further resources
Prompting
Learning Innovation and Lifetime Education (2024). Introduction to AI Prompting (video, 45 minutes). Duke University.
Mollick, Ethan and Lilach (2024). Prompts for Instructors, One More Useful Thing
Generative AI Assignments
Learning Innovation and Lifetime Education (2023). Considerations for AI assignments. Duke University
Delta (2023). Designing Assignments and Activities with ChatGPT and Generative AI in Mind. NC State.
Generative AI Policies
Learning Innovation and Lifetime Education (2023). Guidance on creating AI policies. Duke University
Gladd, Joel (2023). Course policies related to ChatGPT and other AI tools.