So now you’re using generative AI regularly: you have some ideas for assignment designs and you’re incorporating generative AI into your classroom. For example, you would like your students to learn about cavitation.
This all started for me when I watched a YouTube video online. In the video, a scientist hits the top of a bottle with a hammer and the bottom of the bottle shatters. Wait, what? Why? Well, it is not because, as some have erroneously reported, the force applied to the top of the bottle creates compression that breaks the bottom of the bottle. Instead:
If you smack the bottle down really fast, the bottle moves as fast as the hammer and basically leaves the water behind: the water is pulling away from the water and the water stays behind because of inertia. It’s left behind and it creates this low pressure area at the bottom. We filmed at 18,000 frames per second: you can see the cavitation bubbles forming and then you can see them collapsing and as soon as they collapse all the energy at the interface between the vapor bubble and the water gets concentrated into nearly a single tiny point, causing an enormous spike in pressure and temperature releasing a powerful shock wave that shatters out the bottom of the glass.
So I found that really interesting. The only problem: I did not know what cavitation means. So her explanation, without defining cavitation, exacerbated my confusion. How do water bubbles cause an enormous spike in pressure and temperature? So I googled “cavitation” and found the wiki page. The wiki page, in turn, begins with this announcement:
This article needs attention from an expert in Physics. The specific problem is: Several usages of the term appear to be mixed up. See the talk page for details. WikiProject Physics may be able to help recruit an expert. (July 2023)
So this was also unhelpful. The normal solutions were not working! So I turned to generative AI. Research and experience is beginning to show that tutoring is becoming an important use for generative AI in the classroom. Opening Copilot and logging in with my IU credentials, I put some of my prompt engineering experience to work and asked it if it could explain cavitation to me “as if I were a child.” In response, what I received was also useless: Copilot responded by describing cavitation through balloons and party metaphors. Cute, but it also did not clarify for me what cavitation was, how it occurred, and how it was able to shatter a glass bottle.
So I tried again. What would I tell students to do to get a good answer about what cavitation is and how it explains glass bottles shattering? OpenAI offers their suggestions for using prompts for tutoring, but here I thought providing context is the first step to improve my prompt: who is asking the question and how would you like the AI to answer? After a few more trial and error attempts, I adjusted my prompt to:
I watched a YouTube video that says that the reason the bottom of a bottle breaks when you hit the top is because of cavitation. You are a professor in fluid mechanics, but I want you to explain this to me as if I have no knowledge of physics or fluid mechanics (but do not leave any important terms or information out, and do not use metaphors). Be sure to explain how the collapse of the bubbles produced by inertia, which vaporizes the liquid at the bottom of the bottle, creates shockwaves through the heat and pressure created by their collapse.
The response I received to this prompt was something that helped me understand cavitation and then is something I could share with students also. But more importantly, now I better understand how to guide students to use generative AI for tutoring purposes. How can students get answers to questions they have when the instructor is not available? Through coaching students how to use generative AI, we can provide a resource support for them that can help them navigate our coursework. What could you coach students through learning to help them understand how they can use generative AI as a teaching support in your course? An every-growing body of research is pointing towards using generative as a resource for tutoring students, particularly in the realm of second-language learning. Another use of generative AI is to ask it for examples; for instance, ask your favorite generative AI chatbot to give you more examples of cavitation.
To learn more about using generative AI in your classroom read our previous AI blog posts, watch our Faculty Showcase on AI assignments, or contact CITL with questions or for a personal teaching consultation.
Leave a Reply