|
JDV: What would you say are the learning goals for the course? What do you want students to be able to do, to know, or to understand by the end of the semester?
JH: We live in a time now—particularly for freshmen—where everything is changing. And the goal of education is to prepare people, in whatever way, for society, for the workforce, for the world. It’s pretty clear that everything is changing so quickly that in four years when incoming freshmen finish college, AI is going to be something totally different. And so, it seems to me to be almost irresponsible to not actively engage in how AI will be part of what they are going into in four years. It seems like the best way to do that, since I don’t know what is going to be, is to say, “Here’s what I know and here’s how I would use it.” To help people to sort of ride the AI wave—and we will be riding the AI wave whether we like it or not. So that’s kind of what the course is designed to do. The goal of the course is to get students to be fluent in using AI as a tool to help them to be better people. Sometimes AI is useful. Sometimes AI is not useful. We can’t always tell the difference right now. So what I’d like the course to do is to have them compare and contrast different modalities of learning—one of them being using AI, one of them being using books, one of them being using the internet.
JDV: Is your experience that students come to Princeton having used AI?
JH: Oh, absolutely. They have been using it. It’s part of their high school experience. It is absolutely baked into the way they have been approaching learning and articulation and exploration and entertainment and news and everything. It’s baked into the way that they interact with the world because they are a technology generation. They grew up with smartphones. Books and smartphones are both equally old to them! So, since that’s their reality, that’s their ground zero, they’re not coming from a critical point of view. They’re just, you know, human beings living life.
JDV: To get students to critically assess our current moment, you study some of the technological shifts of the past and how they affected society at that time. What is one historical technological advancement that you find to be instructive for understanding the present and possibly the future?
JH: I want what we discuss to be something that students can relate to, and so, I chose two timely books to start the course. One is by Neil Postman. It’s called Amusing Ourselves to Death. It’s about how television, as a technology, really changed society—the entertainment culture versus the fact-based culture. He does a great job analyzing and presenting lots of different perspectives. I start there. Then there’s another book by Jonathan Haidt, called The Anxious Generation, about the impacts of social media. Social media is something that is very much a part of their daily existence—and in teaching the course I learned just how much. Most people in my class subscribe to at least three social media accounts that they’re actively involved with, and they get most of their news through social media.
So I take those examples, and I try to understand how the students have been affected by those technologies. I ask them, “What are your values? What are our values? What is important to us? And how are those values shaped by the things you use? Are they making you more aligned with what you want to be doing, or are they distracting you from it?” And so, we talk about that. It’s actually a perfect time to talk about it because when you come into college it’s a chance to reset and examine the world in a deeper way, and also to self-examine.
JDV: You mentioned that you want your students to “compare and contrast different modalities of learning.” Is there a particular assignment or an in-class activity in which you ask them to do that?
JH: My first assignment is for them to use AI to read a book. One of the things that AI is very good at, broadly speaking, is summarization and doing things like ingesting a lot of material and assimilating it in some way for you. It will produce credible summaries, particularly of short things, though it’s going to be making some choices as it goes along. The goal of this course is to do everything with AI; I want them to use AI, to evaluate it. So, the first thing I ask them to do is to read a book and feel free to use AI. I actually give them the material already digested by NotebookLM (an online AI tool). It produces a short video. It produces a timeline. It produces a podcast which is super-chatty and interesting to hear, as a style. Then I ask them how they felt about it. What did they learn? I ask them to self-assess. Then I ask them to read the book for themselves, carefully, and then again self-assess. Then we talk about it. That gives them real tangible personal experience to use in making their own decisions. And it’s really interesting because they’ve never done both before, so they get a new perspective.
JDV: What were their responses to the two different methods of reading?
JH: They felt like they understood reasonably well with AI. But then when they read the book carefully, they comment, “Oh yeah, there’s a lot about his style that didn’t come through, a lot of really interesting examples. How he built things up made a lot more sense. I didn’t really feel like I got the depth of why he was saying certain things.” It’s an excellent book that has a lot of critical analyses and offers different ways to look at things, but with AI you just get sort of a bullet point summary that just scratches the surface. Then they can answer for themselves “what’s the value of one versus the other?” I’m not saying, “Don’t do this.” I’m not saying, “You can’t.” Do whatever you want! But then let’s see how well it worked. And I think that’s really the value of this course.
JDV: How much time do you spend talking about the technical aspects of the AI tools you’re using?
JH: The first couple classes focus on the society side, and then we start to dive into “how does AI work?” The computer-based classes are in the McGraw Center’s Digital Learning Lab, which has some very capable Mac computers for running AI models—they’re excellent. I had the students actually run different AI models for themselves—ones that they’re not used to using. They can run medium-sized language models to give them a sense of the diversity out there. You know, it’s not just ChatGPT, Claude, and Gemini. There are lots of other models around them and these feed into the larger ecosystem of how society is using AI. I also have them “vibe code” (i.e., have AI generate its own computer programs based on your desired description) to understand how AI can code for you (sometimes effectively, and sometimes not).
JDV: What is the course’s final assessment? Can you talk a little bit about that? It’s a longer paper?
JH: Yes, they are required to do a 15-20 page final paper and presentation on a topic at the intersection of AI and society. They are welcome to use AI to do things, to give them ideas of interesting topics. I took my syllabus, and I put it right into ChatGPT and Claude and I said, “What are some interesting, compelling, final paper topics and titles for this assignment that I could share with my students?” Students will have to choose a topic and make it their own—and if they can’t make it their own, it doesn’t count. I’m grading not just on the paper they produce, but also on the process they used to produce it—how they explored, what their references are, and how they critically evaluate and justify their thoughts. They also need to submit all of their AI interactions so that their research process is transparent. So it’s not just about the product, it’s about how they differentiate themselves. I guess maybe the phrase I want to use is—and I haven’t really said this explicitly, but—“Is it better than AI?” I’d like them to be better than AI.
| 
