This is a post by Dr. Deana A. Rohlinger for the Fall 2023 Newsletter

In Spring 2023, a feeling of existential dread overcame me every time I looked at education news. You probably saw similar headlines:

• ChatGPT Can Get Good Grades. What Should Educators Do about It?
• Paper exams, chatbot bans: Colleges seek to ‘ChatGPT-proof’ assignments
• I’m a Student. You Have No Idea How Much We’re Using ChatGPT

I teach a large online, upper division, writing intensive course titled New Media and Social Change. How was I going to offer an effective online, writing intensive class when students could rely on ChatGPT – and now most search engines – to complete the bulk of their work for them?

I eventually decided that I needed to explore ChatGPT myself. Once I got a good sense of what it could and could not do, I was able create an exercise that would both scratch my pedagogical itch and help students realize that A.I. in its current form could not replace a creative and engaged human. It took more hours than I care to admit, but eventually I came up with the Me and ChatGPT assignment.

I chose a topic from class (in this case, digital technologies and education), asked students to create a free ChatGPT account, and asked them to proceed through the following steps.   

• Prompt ChatGPT to generate a list of how digital technologies potentially help and hurt students in the classroom. 
• Choose two benefits and two pitfalls associated with digital technologies in the classroom from the response provided by ChatGPT. Then, prompt ChatGPT again to provide more information about these benefits and pitfalls.
• Draw on at least three readings from the week as well as two additional peer-reviewed sources and evaluate ChatGPT’s answers, focusing on what ChatGPT did well and where it fell short.
• Finally, use the course readings, your own research, and the ChatGPT responses to write two original paragraphs that explain how digital technologies potentially help and hurt students in the classroom.

The directions students receive are a bit more involved. I provide a template for the assignment and ask students to copy and paste their interactions into the space provided so that I can see how they are prompting the program. Additionally, in their initial evaluation of the ChatGPT answers and of the course and peer-reviewed articles, I allow students to use direct quotes and define what constitutes a paragraph loosely. I focus primarily on their evaluation of ChatGPT as well as the number and quality of the source material leveraged in their analysis.

I also try to help them find peer-reviewed sources. I do this by including a link to a video that reminds them what peer-reviewed sources are as well as a detailed PowerPoint on how they might use course cites to productively leverage ResearchRabbit, an online tool that simplifies the process of literature search, to find additional relevant sources.

My wonderful teaching assistants, Yuki Maynor and Levi Mitzen, and I are grading the assignments now. So far, I’m generally pleased. Students who completed all the steps were able to engage and critique the generic responses provided by ChatGPT and able to use ChatGPT’s pithy responses to help structure their two paragraphs.

Generative AI is still making headlines and still making me a bit anxious. But the Me and ChatGPT assignment feels like a small win. Previously, many of my students were serial quoters, offering random lines of prose from papers that don’t really make their points. This trend, so far, is less apparent in the Me and ChatGPT assignment. They seem to be processing the readings and their sources with more reflection  and, in my opinion, more productively than past semesters.

I will know soon enough if I am right or wrong about my use of ChatGPT in my large online class. I’ve just announced to the class that their final will also require them to use ChatGPT.     

Author Bio

Deana A. Rohlinger is a Professor of Sociology and an Associate Dean at Florida State University.