Much ink has been spilled on topic of college student use of generative AI. Less has been written about how faculty can take advantage off these tools. I’m going to describe a simple application that I’ve actually found useful.
This semester I’m teaching Elements of Data Science, a class that was modeled after Berkeley’s Data 8: Foundations of Data Science, but customized to suit the needs of undergraduate science majors. A series of computer labs assignments are a critical component of this class, labs that many student struggle with as most start with no programming experience. To gauge the pain points, we added feedback questions to the end of each lab:
- How did this lab go?
- Are you confident you could fit a line to a new data set?
- Were there questions you found especially challenging you would like your instructor to review in class?
- How long did the lab take you to complete?
Roughly 200+ students are enrolled each semester across multiple sections of this class, so reading and assimilating all of this feedback for every lab is no small task. This is where AI comes in. I wrote a short Python script to extract the feedback into a text file, which I fed to an AI document summarizer:
- Students found the lab challenging, with some struggling with specific questions and aspects of creating tables in Python.
- The lab took varying amounts of time to complete, ranging from over a week to just over an hour.
- Some students sought outside help, such as referring to previous notes or seeking assistance during office hours.
- Specific questions, like 12, 13, 14, and 17, were highlighted as particularly challenging by multiple students.
- Despite the challenges, many students found the lab rewarding and helpful in their journey to become proficient coders.
I found this quite useful, particularly the list of questions multiple students found challenging. Maybe generative AI’s impact on college instruction won’t be all bad.