by Robert Cole, Program Director, Reinert Center
In a recent workshop I was asked about how to have a conversation with a student suspected of using generative AI to complete an assignment without permission to do so. I’ve been thinking a lot about this lately as we all continue to move through this time of generative AI novelty and how it is affecting our classes. Even I have caught myself thinking, “Wow, this is pretty good work from this student… I wonder if they had some AI help”. Upon further reflection I realized I had probably done a disservice to both the student that worked really hard to produce the work and myself for not having more faith in my student’s abilities. If I had suspected my student had used generative AI, flouting my guidelines to the contrary, what might be a good way to approach the conversation?
We can begin this conversation very early in the semester, perhaps the first day. Consider discussing academic integrity as early as possible in the semester. Even if you have thought that others have probably talked about academic integrity earlier in a program or course of study, think about discussing it in the context of your class and potential professions. You may be able to explain that integrity is not only an expectation in your classroom, but also an expectation in students’ profession of choice. Lack of integrity in the profession may well have a greater impact than a bad grade. Think about your position around the use of generative AI and make sure students understand your expectations – make them explicit. Providing your thoughts around not only what generative AI can do, but also the limitations it has relative to your context can help generate a sense of value that may mitigate cheating. Finally, you can help students understand that you are literally there to help them learn and will support them in that endeavor. Think about the alternatives you can provide or resources to which students can be directed. You can offer office hours, scaffolded content, referrals to the writing center or to campus tutoring.
Something we may need to keep in mind are the contextual factors that may have led to any academic integrity infraction. While not necessarily excuses releasing students from responsibility, the context may have pertinent baring on how we handle the suspected infraction. Students are more likely to seek help from generative AI – among other helps that are not permitted – if they perceive they have a lack of understanding of the content to successfully complete the assignment. In addition, within an arguably competitive system, fear of failure may push students to seek what they perceive as an advantage by employing generative AI. When so many things seem to be vying for our time, if students perceive they do not have the time to complete an assignment by a very firm due date, they may reach for a way to complete it much more quickly than they would otherwise be able to, with a tool like generative AI for example.
Once you have in mind what you have explained to students regarding generative AI earlier in the semester, and the perceptions they may have regarding their current situation within the semester, you can turn your attention to the conversation you want to have. Consider planning the conversation to the extent you can. Try to remember that the infraction you suspect is not a reflection of or an attack on you or your course (Lang, 2013). It is most likely a reaction to where students feel they are in the moment. Knowing these things, think about framing the conversation more positively than immediately making bold accusations – that you may not be able to prove. Concentrate on the student’s work; things you found interesting and things you have questions about. Try to avoid assuming a student has cheated. Asking questions to allow a student to explain an approach to the assignment may prove very enlightening and might reveal legitimate methods you hadn’t considered. Questions like, “I noticed some interesting elements in your writing… Can you tell me how you developed those ideas?” may be helpful, too. If you have spent time forming relationships with your students, these conversations may be more productive and less stressful for both parties.
If further information is needed to make a determination of what may have happened, consider pointing out inconsistencies in style and/or content. You could ask things like, “I’m curious about this section here… it seems a bit different from the rest of your writing. Can you explain how you came up with this?”. Questions related to sources may be helpful as well. You might ask, “Where did you get the information to support this claim?” or “How did your find this source related to…?”. In addition to the previously mentioned strategies, some impromptu revision may help you see how a student is thinking. For example, asking, “I’d like you to revise this section, focusing on strengthening the argument with your own analysis. Can you walk me through how you would do that now?”
While this approach will not ensure a confession of wrongdoing, with a more educative approach the message to the student is that you care about the learning that is expected, not just how the artifact should be created. Opening a dialogue centered on learning and empowering students can be far more valuable in the long run than simply punishing a suspected transgression.
Lang, J. M. (2016). Small teaching: Everyday lessons from the science of learning (First edition). San Francisco, CA: Jossey-Bass.
The Notebook 