If I Don’t Talk About GenAI, Does it Really Exist?: Introducing GenAI Using a Cover Letter Assignment DOWNLOAD PDF

Meredith Geller

Director of the Writing Lab and Clinical Professor of Law

Northwestern University Pritzker College of Law

Published: December 2025

If you are anything like me, you know you need to incorporate GenAI into your class but aren’t sure how to do it. You want to walk a line between acknowledging GenAI exists and not being so accepting that your students feel they have your approval to use GenAI for everything instead of doing work on their own. Most importantly, you’re worried because you’re far from an expert and aren’t entirely sure how to explain GenAI to a classroom full of students who may understand it better than you do. 

First of all, do not panic. GenAI can sense your fear. 

No, really, do not panic. Like teaching any new subject, you want to educate yourself as best you can, but you do not need to be an expert to start. I knew I needed to introduce GenAI into my classroom, but I was an English major with zero computing background. To educate myself, I read some of the scholarship from our fabulous colleagues^[1] and began playing around with Lexis Protégé^[2] to see what options my students had and how I could use that in class. Although I have used ChatGPT, I especially like Protégé because it always links the sources it references, which means that I can go directly to the source material and have fewer worries about hallucinations. Protégé also saves your searches and results, so I can easily go back and find prompts that worked well and those that didn’t.

To further educate myself, I took a GenAI course offered by my university. The most valuable part of the course for me was an empathy map that we created to address student concerns about GenAI. I realized while filling it out that I had no idea how my students felt about GenAI. As a result of that realization, I asked the students in my upper-level Writing for Practice class what they thought about GenAI. It turns out that my students had the same concerns about GenAI that I did as a professor. Students thought it could be used to assist them with a variety of skills and topics, but worried that it was taking the place of critical thinking and that some people relied on it too heavily. Students shared experiences with friends in other graduate programs who took classes for which they did no homework, instead relying on ChatGPT to answer any questions their professor gave them. They were concerned about how this overreliance on GenAI would impact service-oriented professions, like social work and teaching, as well as the law. Students also stated that drafting a decent prompt can be so time-consuming as to discourage them from using GenAI since it’s easier and faster to just perform the task themselves. Finally, many students were incredibly concerned about the environmental impact of GenAI because of the resources it consumes. Because of these concerns, some students avoided using GenAI entirely. 

Although I had already planned to teach a class session on GenAI—despite being terrified to do so—and give my students an in-class GenAI assignment, I used this discussion as a foundation for that class. I also applied my experience with Protégé and what I had learned in my reading and in the university course. Most importantly for me, I told the students up front that I was far from an expert and that it was entirely possible that some of them knew more than I did. Despite my fears, the class went well. We talked about the importance of prompt engineering, and I walked the students through some prompts I had drafted in Protégé that worked both well and poorly. We discussed hallucinated information and the importance of checking information received from any GenAI model against reliable sources. Finally, we discussed the importance of retaining and using critical thinking skills when assessing results from GenAI.

After this discussion, I gave the students an in-class assignment for completion points only. I wanted to provide the students an opportunity to play around with GenAI like I had in a stress-free environment. As part of the assignment, I gave my students a prompt that would encourage them to create a short, simple document with which they were familiar, a cover letter for a job. The prompt I gave my students was:

Ask Protégé or CoPilot or ChatGPT or an AI model of your choice to write a cover letter to a potential employer. Feel free to give the model as many or as few details as you’d like. Once you’ve gotten that cover letter, review it and let me know what you think about it. Is it something you would use? Would it be helpful as a starting point to you? Briefly explain why or why not?

The feedback from my students was very informative. Students were split about the helpfulness of the cover letter GenAI drafted, with more leaning toward the negative. Many students felt the GenAI-drafted letter was robotic and disingenuous. These students noted that even with prompting, the results often sounded bland and so general that it was obvious GenAI had been used to assist in the drafting. They also felt that the GenAI-drafted letters lacked a personal character that would lead those in charge of law firm hiring to immediately know that the letter had not been drafted by a person. One student also reported experiencing hallucinatory results when the model they used created a summary of their law review article that matched neither the article title nor topic.

In contrast, some students thought the drafted letter was a good start and could help while facing the dreaded blank computer screen. These students reported that they often struggled coming up with different ways to express themselves in cover letters, and the GenAI model gave them some suggestions to move forward. However, even these students acknowledged that it would just be a draft and/or a template since the results did not sound personal enough. Since the point of a cover letter is to set yourself apart from the crowd, the more generic result that they received, the less likely a GenAI-drafted cover letter would be useful to them.

Even though it forced me out of my comfort zone and required my learning enough about GenAI to convey to my students, I really liked both the class and the exercise, and so did the students. For students who were interested in GenAI, it gave them the opportunity to try new things, including practicing prompt engineering for a familiar task. Even though I did this assignment with upper-level students, this assignment would work with 1Ls as well and may even be more useful to them, especially in the first semester before they send out letters and resumes for summer jobs. As 1L students may not have ever drafted a cover letter, this exercise could be an especially helpful way to discuss the strengths and weaknesses in AI-generated documents. Regardless of when it’s used, this was a simple, short, and illuminating exercise that I was glad to have shared with my students.