What can I do if I suspect a student has cheated with AI?
It can happen to the best instructors: you begin reading through your students’ writing and your heart sinks. Some responses sound canned and strangely superficial. A couple of the assignments seem to have the same recycled examples. A student who struggles with writing has turned in work that is written in a very different “voice.” You begin to suspect that some students have used AI to do some or most of their writing.
One (very natural) reaction to this suspicion is frustration and anger. It is certainly upsetting when the course (and discipline!) that we care so much about seems to be so low on a student’s set of priorities that they don’t even bother to engage with the meaningful assignments we create. The thought that a student simply offloaded their responsibilities to a technology like AI is upsetting. These feelings may lead us to simply put the suspect paper through an AI detection software, see a high probability that the text was generated by AI, slap a failing grade on the assignment, and move on.
But there is another way! While having an emotional reaction is normal, it is important to move beyond that initial reaction and respond to our students systematically and logically. In this short guide, we suggest a series of steps that you can take to see this situation (and your students) with greater clarity and discover surprisingly simple and highly productive ways to respond to suspicions of cheating with AI. Read on!
Don't rely on AI detection software.
Using AI detection software has two big drawbacks: one is a technical drawback and the other is a relational drawback. The technical drawback is that this software is notoriously ineffective. AI detection software is frequently wrong, generating false positives. In addition, when we allow our frustration to dictate our response and focus on “detection,” we immediately create roles for ourselves and our student: the punisher tries to root out the criminal. Putting ourselves in this punitive role makes it difficult to really understand what our students may be doing when they lean into AI and makes it harder for us to build and maintain a relationship of mutual trust with them.
Consider students’ lives and perspectives.
Students who use AI inappropriately aren’t necessarily unethical or immoral. In fact, the research on cheating suggests that academic dishonesty can arise when ethical students don’t see meaning in a course, fear failure, or are overloaded. Our students are frequently managing a lot of course work, jobs outside of the university, and maintaining their health and social life. In short, their lives are complex.
The rapid proliferation and availability of generative AI tools has introduced additional complexity, which is amplified by the fact that students are often receiving different messages about how they should (or shouldn’t) use AI. For example, instructors in different courses often take divergent approaches to AI. Some faculty want students to use it, and some don’t; some give careful guidance about AI in relation to schoolwork, and some give very little. It’s also good to keep in mind that well before they get to college, students may have been encouraged to see AI as a “tool” for learning. When we remember that our students are still developing as learners and as people, we realize that navigating AI use in the face of these competing demands and messages is particularly challenging for them. As a result, it is possible that our students may use AI even when we have explicitly asked them not to.
Give the student the benefit of the doubt.
Even when we have asked students not to use AI to complete work, they might believe that they are using AI in appropriate ways. For example, we might ask them to compose their own short responses to an essay question. But what if a student feels their original writing seems disorganized? Perhaps that student uploads their essay and asks AI to fix grammatical and organizational errors. Is that cheating? While you may think it is in your course, consider that it might be acceptable in other courses your student is taking.
Not only should we thoughtfully ask ourselves if students intended to cheat and not do the work we wanted them to do, but we should also remember that our students are still learning to research, to write, to solve problems, and to use the web ethically. This means that they may be struggling not only with appropriate use and attribution of sources, but also with writing in their own voice. If they are conducting web research or using AI in appropriate ways, they may begin to absorb that voice and inadvertently incorporate it into their own writing. In addition, novice writing has many of the same characteristics as AI writing: it can be flat, impersonal, and lacking in variety. Keeping this in mind when we read students’ work can help us slow down and not assume that students set out to deceive us.
Set up a time to meet with the student.
After we consider our students’ experiences and their emerging understanding of academic work in our disciplines, the logical next step is to simply meet with a student we suspect may have used AI. Keep in mind that this conference with the student should be held to learn more, not to accuse.
How should you set up this meeting? Be honest and kind. Ask the student if they have a moment to talk at the end of a class and ask them to set up a time to meet with you. Your invitation to meet should go something like this: “I would like to talk to you about your [paper / homework / project]. I want to make sure you feel supported in the course on these kinds of assignments. Could you come to office hours this week?” If the student asks why, you can respond, “I’m a little concerned that my guidance on what resources to use is not clear. Your work is good, but I want to make sure I’m really hearing your voice / seeing your work. I’m worried that you might be leaning into AI or other resources and if that is the case, I’d like to help you resist that. Our meeting isn’t about me punishing you, but rather about making sure you can succeed.”
If you think that having this kind of conversation could be unnerving to the student, put your invitation into an email (using the same kind of language) and make sure to follow up if you don’t get a response right away.
Meet with the student and let them lead.
When you meet with your student, keep in mind that they will probably be feeling worried and defensive. Start the meeting by reminding them that you want to help them learn and succeed and that you invited them to meet so that you can learn more about their work and their thinking. Ask them how they are doing in the course and then ask them to describe how they worked on the assignment. It is not unusual for students at this point to share where they are struggling or to admit that used a work around to offload some of the work of the assignment.
How you respond to an admission of cheating can make or break your relationship with that student and can also be a turning point in terms of their learning. If a student admits they cheated, find out why. Are they confused? Are they strapped for time? Are they struggling in school or outside of school? Respond by helping the student make a plan to manage the challenges they are facing, keeping in mind the resources you’ve built into your course but also the campus resources that are available to them. Some things you can do include the following: help the student plan (in writing and with a calendar/planner!) how they will tackle the next assignment; set up regular meetings with you to review drafts or homework assignments in the coming weeks before they are turned in; or support the student as they make plans to meet with tutors, library staff, or their advisor for more learning support.
It's also important to consider how you will respond if your student denies that they have used AI and you are fairly sure they have. One option is to ask, without accusing, that your student share drafts, provide access to a shared document with a history of the writing, or share notes and calculations that show how their thinking developed as they did their work. Another response is to point out the ways in which the writing or the work the student turned in falls short of your expectations. If a student did use AI, this will not only help them see how unproductive cheating is but will also help them understand your expectations more fully.
Require the student to rework the assignment if they did use AI inappropriately.
If together, you and the student agree that they relinquished some of their learning to AI, remind them that your job is to teach them and that you are concerned that they cheated themselves out of the learning that the assignment was intended to help them do. Make sure that your student understands how they should or should not use AI on the assignment in question and on future assignments. Then, rather than give them a failing grade, require them to rewrite the assignment. This step is crucial and is more rigorous and demanding than simply giving them a failing grade for the assignment. It also means that you are helping ensure that students have the opportunity to do the learning the assignment required. Students might be able to earn only 70% of the original grade. Alternatively, you may decide that the resubmission can earn full credit.
Help them move forward with support and structure.
When students cheat, they usually do so because they are unprepared. To help ensure that they will engage in the next assignment in a more productive way, it is crucial that your student leave with a clear sense of how to prepare. You have already spent some time helping them plan and consider strategies and supports that can guide them toward the assignment, and it is important to talk about what they should do if they find themselves backed into a corner and tempted to use AI to do their work. Talk through alternatives such as coming to you and asking for an extension. Also discuss clearly what you will do if the student does cheat again. This is something you will need to think through ahead of time: carefully consider what you will do if a student cheats a second or a third time. For example, you may require the student to redo the work the second time for fewer points and the third time they will get no credit at all. Develop your ideas about consequences of subsequent incidences of cheating with support from colleagues and CATLOE, and be sure that these consequences are in place to help students regulate their learning rather to simply punish.
Draw on the experiences of student academic dishonesty to develop as a teacher.
Each time a student cheats, we have an opportunity to learn more about our students and our teaching. This is a very different perspective from that of instructors who feel their job is to catch cheaters. In fact, it is the perspective of a reflective, curious academic and one that drives your scholarly work. By taking the steps we’ve described above, you will doubtless learn more about the students in your classes and recognize that they experience your course quite differently than you envision it, especially when they face meaningful and complex challenges. This may mean that we need to rethink how we design the phases of learning in our courses: our students may need to work toward assignments, exams, and project in stages so that they get more low-stakes practice and build up more strength before they face high stakes assessments.
Additionally, when students cheat, you may become aware of the need to design motivating course policies that help you and your students think through the value of engaging with assignments honestly. These policies can be a springboard for important conversations about the role of AI in learning, in your course, and in your discipline. The research suggests that students want guidance about AI use and ethics in college. The student who cheats can help you learn why students use AI unethically and can be a guide to helping you help future students resist that temptation.
Resources
- Goebel, C., Strauss, D., & Tessier, N. (2024, May). Ai and Academia: Student perspectives and ethical implications. StudentPOLL Art & Science Group, 17(1).
- Lang, J. M. (2013). Cheating lessons: Learning from academic dishonesty. Harvard University Press.
- Lang, J. M. (2020). Distracted: Why students can't focus and what you can do about it. Basic Books.
- McCabe, D. L., Butterfield, K. D., & Treviño, L. K. (2017). Cheating in college: Why students do it and what educators can do about it. Johns Hopkins University Press.