Fred Schaffer, Professor of Political Science, discusses how he developed an AI policy for his Integrative Experience course that connects his course learning goals to his choices about AI use in his class.
What initially prompted you to develop this policy?
I teach an Integrative Experience (IE) course, and spent the summer before last revamping it with the help of Colleen Kuusinen from the CTL to directly link each aspect of the class to the course IE learning goals. To give just one example, I have students write short preparatory assignments before most meetings, in which I ask them questions about the assigned readings or videos. After working with Colleen, I began indicating exactly what IE learning goals each preparatory assignment was designed to advance.
That worked pretty well, but with the advent of AI, I began to suspect that some students were using ChatGPT or other AI tools to complete their assignments. I decided to write up a detailed AI policy, taking as my starting point the lessons I had learned from Colleen about linking each element of the course to the goals it’s intended to achieve. I didn't only want to tell students what was allowed and what wasn't allowed. I wanted to articulate as clearly as possible why the rules I adopted helped further the learning goals of the course. I wanted students to understand how the policy was designed to help them develop skills that would benefit them in their lives beyond college.
How did you design your AI statement?
I created a document that indicated what was allowed and what was not, along with the rationale behind each decision. I only allowed grammar checking and minor edits using tools embedded in Microsoft Word, Google Docs, and Apple Pages. Students were allowed to use these tools to fix minor and isolated grammar and spelling problems. My rationale was that if students paid attention to what they're fixing repeatedly, it would help them improve the mechanics of their writing.
There were four categories of things not allowed:
- Rewriting and restructuring entire paragraphs — that would hinder them from developing better communication and critical thinking skills. Putting ideas together in a coherent way is a thinking skill just as much as it is a writing skill.
- Outlining or drafting entire assignments — This would simultaneously undermine all the learning goals for the course.
- Using AI to generate ideas — Letting a robot do their thinking for them would short-circuit the kind of thinking skills that hope students develop.
- Summarizing the readings — This would undermine the content and comprehension learning goals of the course.
For every writing assignment, I also required students to provide an assistance disclosure. I asked them to indicate what grammar checker they used and to give a sentence or two about the types of errors, problems, or misspellings that were caught.
Additionally, at the end of the semester, I gave students an extra-credit opportunity to review their assignments and disclosures, reflect on the strategies that they used or developed over the course of the semester to improve the mechanics of their writing, and identify two or three goals to improve their writing moving forward.
How do you talk to your students about your AI policy?
This past semester, I devoted a class meeting to the AI policy. I assigned a New York Times podcast in which a journalist interviewed a bunch of college students about their AI use—a helpful learning tool because it doesn't give just a single perspective. Then we had a class discussion, focused on a single question: Why bother writing something that a robot could do?
This was an important discussion to have because it's so easy to use AI. Students have to see for themselves why putting in the hard work of developing reading, writing, and thinking skills is worth it. What do you bring to the table, either for an employer, in graduate school, or whatever? What is it that you can do or contribute that a bot can't? I wanted students to think about these questions and talk about them together. Only after having that discussion did I introduce my course AI policy.
What impacts have you seen since adopting this policy?
I noticed towards the end of the semester that students started adding comments to their assistance disclosures. They would write things like “I didn't use the grammar checker so much on this assignment,” or “It’s finding fewer mistakes,” or “I'm learning how to identify the kinds of errors I make, and I'm starting to be able to look out for them on my own.” I think the policy helped them improve their writing skills, build their confidence, and give them a sense of pride.
In terms of the extra-credit reflection opportunity, I think about a quarter of the class did it. I was struck at how several students noted that they had never examined the mechanics of their writing before. This was a new experience for them. Overall, though, I was impressed. Different students had different sorts of writing challenges and came up with different sorts of writing plans. I hope they follow through.
Is there anything you are considering changing?
This past semester, the final reflection was for extra credit, but I may integrate it as a required component of the course moving forward. I take student feedback seriously, so I won't decide until I get the SRTIs back. The final reflection adds yet another layer of work for students, and I want to make sure that there's a sufficient payoff to make it a requirement. For what it’s worth, I did give anonymous surveys during the semester about things that were working or not working for them, and the assistance disclosures did not come up much as something that students found onerous.
Taking a step back, I readily acknowledge that in certain courses with different kinds of learning goals, there could be good reasons for using AI, but for the basic learning goals of my class—learning how to read carefully, write clearly, and reflect critically—I have found so far that AI hinders more than helps. But my thinking on AI is evolving, and I’m open to exploring in the future how it might be used in a targeted and delimited way to promote the learning goals of the course.
Any advice to instructors as they consider a statement of their own?
I found it helpful to revisit the work I had done with Colleen about aligning course assignments and policies with its learning goals. That provided me with structure for thinking about the AI policy that I wanted to create. I imagine that sort of structured thinking would aid other instructors as well. I believe it is important, too, to emphasize the payoff for students. Rather than just laying out expectations, it helps to focus on what’s in it for them.
