Last semester, I suspected I had a major issue with the use of generative AI in my survey courses, so I inserted what is known as a “Trojan horse” into the directions for a paper I assigned. As it turned out, I did, in fact, have a major problem, and my post on Threads about it accidentally went quasi-viral and ultimately became this HuffPost article, followed by an NPR interview. Fellow educators and the majority of other readers were generally horrified by the results of my experiment, while a minority accused me of setting my students up for failure and creating a classroom atmosphere of distrust — as if students’ use of generative AI had not already done so.The results of my approach, though, raised a serious question I have tried to answer: How do I de-incentivize AI use? I spent the past semester experimenting, and while I don’t have all of the answers to that question, I think I do have some. First, I reconsidered my assignments. Instead of the weekly essay that asked students to write an analysis of secondary and primary sources, I pivoted in part to role-play essays in which students used information from the lecture and assigned reading to craft a response from the perspective of a character they created.For example, in one assignment, I asked students to imagine themselves as a Gilded Age laborer — they could choose a U.S.-born laborer, immigrant, or child — and write a letter or diary entry from that perspective that explained their day-to-day life, challenges, fears and hopes. Another assignment asked them to write as an average citizen accused of being a communist by Joseph McCarthy. They had to grapple with the accusations, the repercussions and the global context. In essence, they had to write as victims of social hysteria.Some students wrote truly moving responses and clearly committed serious time, imagination and even emotion, while others simply did what they needed to do to satisfy the requirements. But the key results were two things: They did the writing, not AI, and some made a sincere attempt to see an issue from someone else’s perspective. They tried to write with empathy.Second, I changed my semester-long project from one final turn-in into a series of small checkpoint submissions. This allowed me an opportunity to check their progress and see their thinking unfold over time. This did not work as smoothly as I would have liked, but it did help mitigate the possibility of getting an entirely AI-generated final product, which has happened more than once. Besides scaffolding the semester project, though, I also considerably altered it. Students have always had three options: a web-based digital presentation, a short video presentation or a final paper on a topic that I assign. I removed the final paper option and replaced it with a physical art option, inspired by a former student who once produced a great piece of art on ping-pong diplomacy.Only a handful of people chose the art project, but something happened that I did not anticipate — nearly every person who produced a piece of art chose to depict historical moments in either the women’s rights movement or the gay rights movement. Consider that I am at a small regional university in the Texas Tech system, where new censorship rules have drawn national attention, and the students’ decisions are all the more striking. But those students seem to know a basic fact that is often lost on many: Real art does something generative AI cannot do. It speaks of the human experience. It pushes boundaries, expresses anxieties and fears, and challenges an unjust system. They weren’t required to do these things, but when given an option, those students made a choice to do them. “I didn’t just ban generative AI without telling them why ... I also made it known that they had a choice, and that choices have consequences both inside and outside the classroom.”Role-play and art assignments are not new, but, finally, in what I think was the most important shift I made, I changed my approach during the first week of the semester. The first day of class — otherwise known as syllabus day — became syllabus week. Instead of just showing them a list of course policies, specifically a ban on generative AI use, I asked them to consider why that ban existed. I assigned an essay by Patrick Lin at Cal Poly that explores the pros and cons of generative AI, asked them to write a short response to it and then had them discuss it in class. I let them direct the discussion after a simple question: What about Professor Lin’s essay surprised you the most? The responses were varied, but they seemed to focus on three main points. First, that studies show that the excessive use of generative AI creates serious cognitive and psychological risks. Second, that AI can be wrong and will work to placate you. Third, the environmental impact.I then asked them a few questions. How many of you grew up in a classroom with a laptop or tablet? The response was everyone. Did you all know that we now know that this was a mistake and learning better takes place with physical books, paper, and the act of handwriting? No one knew that.How many of you went remote during covid? Everyone. Did you all know that decision caused serious gaps in your education and development? That you know less than we expect you to know about both factual information and being a student? Did you know most of you are not prepared to be here? No one knew that — this also got some expressions of horror. Finally, did you all know that data suggests Gen Z is the first generation in over a century to test lower in intelligence than the generation before it? This got a serious response of shock and some laughter. But it definitely got their attention. (Jared Cooney Horvath had just testified about this in front of Congress a week earlier when we were having this conversation.)I then asked them one more question: Why are we discussing generative AI in a history class? There were a few attempts at an answer, but most of them were generally unsure yet curious. My answer came out a bit differently each time, but below is the explanation I gave them.“What you will learn throughout the semester is that history only moves forward when people collect information, think about it deeply, decide what it means to them, and act upon it. Our collective national history is complicated, but the decisions of all those who came before us created the nation in which we now live. Rights, democracy, freedom…the things we speak of so freely do not exist just because they do. They are consequences of reading, writing, thinking, and then acting.” “If we outsource our opportunities, our education and our minds to generative AI, not only can we not build something better, but we risk losing the progress made.”“We inherit their victories and their failures, and it is our responsibility as citizens and as people to learn all that we can to avoid those same failures and to achieve new victories in pursuit of creating a better world – one that we may not live to enjoy, but that we leave for those that follow. If we outsource our opportunities, our education and our minds to generative AI, not only can we not build something better, but we risk losing the progress made.” “So, we talk about generative AI in a history class because history is filled with people that made decisions. When it came to laptops in the classroom, you had no choice. When it came to COVID and lockdowns, none of us had a choice. You were all children and had no say. You trusted the adults in the room to do the right thing, and now you must grapple with the consequences.”“But you are no longer children without a choice. Like the historical actors that came before you, you have a decision to make. You are the laptop generation. You are the COVID generation. Will you be the generative AI generation? That is your choice. Will you let yet another thing undermine your capacity to learn, to grow, to think and to build? Or will you decide to take the other path? This is where you write your own history.”The response was always silence — maybe they were considering what I said, maybe they thought I was crazy. Maybe it was a mixture. But they definitely listened. Was my approach perfect? No. Was it successful? In part. Did it entirely remove the problem? Absolutely not. But as an educator, all I know how to do is educate. As an historian, all I know how to do is understand the present and contemplate the future by examining the past. So, I educated them and asked them to take a moment in their present to contemplate their future. I didn’t just ban generative AI without telling them why, and I think that helped with rapport. I asked them to trust the process, and many of them did. I also made it known that they had a choice, and that choices have consequences both inside and outside the classroom. These are not all of the changes I have made. Like many others, I’ve gone back to blue books. But going forward, for now, I am going to stick with this approach and continue to develop it. Giving students choices and explanations seems to help, as does giving them freedom to imagine, create and speak by removing the rigidity of more typical essay assignments. This semester, I received some of the most positive in-person feedback from students I’ve ever gotten. I was thanked for my honesty, for being direct with them on difficult topics and for allowing space for their self-expression. So, I guess my advice, if there is any, is to take a moment to educate students about the harm generative AI does and then let them make the decision about how to go forward. Show them that you care, and they might care, too.Do you have a compelling personal story you’d like to see published on HuffPost? Find out what we’re looking for here and send us a pitch at pitch@huffpost.com.Relatedcollegeartificial intelligenceHistorycheatinghigher education