AI Plagiarism Considerations Part 2: When Students Use AI
The contours of the conversations I want to have with students about their GenAI usage
In the last post (which got a lot more attention than I thought it would—welcome new readers!!!), we talked about how to navigate the AI plagiarism checker conversation and why. In this one, I’m going to provide some guidance and insights about how faculty might engage with students once they suspect that a student might have inappropriately used generative AI. (You can read part 3 of this series here).
Caveat #1: Before starting this, I do want to unpack my own biases in this conversation so folks understand where I come from when I step into this conversation. I know many folks are in the same space but others aren’t and I think it’s important to realize that before seeing what I recommend.
My take on plagiarism and “academic dishonesty” has evolved over my 18 years of teaching from seeing it as the “capital crime of academia” (yeah, I used those terms) to believing it is something to understand but not necessarily worry about. There’s a longer write-up about that journey but for now, I’ll just say this: there are big and small reasons it happens. At the end of the day, it happens because trust was not established or maintained in some way. And the first thing I need to do is figure out why.
Additionally, I think there’s vast “dishonesty” perpetuated throughout a university in the student’s experience and learning including an institution’s own “academic dishonesty”, and I can certainly understand why students are likely to end up participating in such practices. Not saying that it is exactly tit for tat, but what students do is a microcosm of the institution, which is its own microcosm of the culture at large.
That doesn’t mean I don’t at times still go through the emotional rollercoaster when I discover a student has misrepresented their work or their efforts, but I step back and recognize that a good part of that rollercoaster is ego. Yes, there are objective reasons for concern about “academic dishonesty”, and yet, I do not feel they require a deficit lens to address or resolve. Again—way more to say about this but maybe that’s another post at another time.
Caveat #2: I’m focusing this on traditional assignments. I’m also less and less and less a fan of these assignments. That was before generative AI and doubly so now. So this shouldn’t be read as me endorsing traditional assignments like essays and research papers wholeheartedly but rather if you use those, consider the ideas herein.
What If A Student Says They Used AI
In the last post, I mentioned that a conversation is needed—it doesn’t matter how much “proof” is accumulated; the conversation is important. I have seen (and been) the instructor who just inserted the 0 for (traditional) plagiarism, explained why, provided proof, and went along with the day. It’s tempting to do and yet, missing an opportunity to learn and improve for both the student and instructor.
Once you have engaged the student about their inappropriate use of generative AI, if the conversation produces a moment with a student saying they did use it, that’s really helpful. Some students will be that straightforward because they don’t see it as a big deal, they didn’t fully realize it was a concern, or their anxiety is up a few notches and just want to be done with it.
At this juncture, I’m going to thank the student for letting me know and then ask them if I can ask them about their usage. I’m interested in learning more. I want to move the discussion into an interview about their usage to maximize my own understanding. As importantly, it becomes a bit of a reflection process for the student that can also yield results for them.
When a student shares that they used it, we want to avoid the “ah ha—gotcha” or “I knew it!”. The question to consider is if they have used it in a way that violates your course policy or the institution’s policy. One thing I’ve learned is that while policies in the course and institution are helpful for guidance, they also do not necessarily cover all the contingencies and students are going to come up with really interesting use cases that I want to learn more about to help me also understand how generative AI tools might differently support or interfere with learning. This is also why the interview is important—you don’t know yet if their use violated the policy.
After learning from the student how they did it, you want to figure out why. The real meaning-making begins with finding out why they used generative AI. I want to deeply understand why and how the student ended up using generative AI. Yes, there are going to be surface reasons but go deeper to learn more. We still aren’t great in understanding the moments and decision points of when and where students veer off from doing the learning and work we are asking of them. The more we can understand those crossroads, the more we’re going to be able to anticipate, support, and provide alternative pathways that work.
And I know that I’m never going to achieve 100% of students doing assignments the “right way”. That’s not the purpose for me or even realistic. I am interested in finding meaningful ways that students can plug into learning activities and feel like they better understand when and where they may need to call upon help. It’s for those moments that I want to have interventions planned or queued up for them.
I’m always going to be interested in that why and how they got there. Therefore, I’m always going to encourage learning more from the student and doing so in a supportive way. So it’s more than just asking why they did it if they share that they did but getting into the details. The following is a set of questions you can use to help surface more details.
Were there any time constraints or external pressures that influenced your decision to use AI?
How do you usually handle challenging assignments or tasks without using AI?
What support systems or resources do you currently use for your studies?
At what point did you decide to use AI?
Why did you turn to AI and not a peer or the instructor?
Did you try any other resources or strategies before using AI? If so, what were they and why did they not work for you?
Can you describe how you approached the assignment before deciding to use AI?
Why was this assignment different to lead you to AI?
What was going on in your head when you did?
What specific aspects of the assignment did you feel least confident about?
What would have empowered you to move through the struggle?
How do you think using AI impacted your understanding of the assignment's content?
What benefits do you think AI provided that you couldn't find elsewhere?
What did you learn by using AI? How do you perceive the role of AI in your learning process?
What might you be missing or not able to do as a result of using AI?
If given the opportunity to redo the assignment without AI, what would you do differently?
What feedback or guidance do you think would be most helpful from instructors regarding the use of AI in assignments like this?
BTW, I started the list above and then used those questions to get more questions from generative AI, which then helped me come up with additional questions. You can check out the chatlog here.
Again, I want to make sure to ask these questions with curiousity and not like it was an interrogation.
Now, where does the conversation go after you interview the student. That’s harder for me to advise because I think though the interview you’re going to end up with a lot more to consider. My tendency is going to discuss with the student where they went astray and figure out what would be a good means to split the difference or most important, achieve the outcomes that were intended by the assignment. But that would be in conversation with the student based upon what I learned.
This might mean suggesting they do it again from scratch or some other path. It would also include clarifying how they should proceed without (or with) generative AI to complete it.
What If A Student Says They Didn’t Use AI
All of the above is easier to do if the students shared that they used Gen AI or some other indicator of additional appropriate or inappropriate help. Yet, what if an introductory conversation about noticing differences in their work does not yield them sharing their usage? This is much trickier.
It’s trickier in part because it means (at least for me) that (if in fact they did use Gen AI) I haven’t established enough trust for them to share that they did. I haven’t disarmed the fear bomb as it were.
It’s also trickier because they may not have in fact used gen AI and therefore, the more I push, the more trust I lose and the more harm I cause. I think about this a lot because in my own past, I’ve been accused of things I did not do and also, could not prove that I did not do. It’s not a good feeling at all and it erodes one’s trust and faith in the space. It’s a hard slap in the face than when you’re doing what you’re supposed to and you’re told otherwise. It’s academic gaslighting at its worst.
What to do? Proceed with caution. Again, this is an opportunity ask questions and learn. No matter what, you and the student will learn more about the work and their learning. In truth, these are largely good questions we can ask all students as part of a reflective practice.
Can you walk me through your process as a whole or around this portion of the work?
What did you learn about this process that you would want others to know or try?
Can you elaborate on how you developed your thesis statement/main ideas/central theme?
Which part of the paper did you find most challenging to write, and how did you overcome it?
What part of the writing process did you enjoy the most, and why?
Can you discuss a specific argument or point in your paper that you feel particularly passionate or proud about?
How did writing this paper change or reinforce your views on the topic?
How did you end up deciding about [this point]?
Why did you find it valuable to include [this point] in the work?
How did you organize your ideas before writing?
What was your strategy for integrating sources into your work?
What source(s) did you find most useful or insightful in your work and why?
Can you describe a particular source that significantly influenced your perspective?
I know citation styles can be tedious; I know I struggle. How did you ensure your paper adhered to the required citation style?
How did you approach the revision process for this paper?
Can you give an example of a change you made during revision and explain why you made it?
What feedback did you receive from peers or instructors, and how did you incorporate it?
What new skills or knowledge did you gain from this assignment?
Can you explain the connection between this paper and what we’ve discussed in class?
If you had more time to work on this paper, what additional research or revisions would you consider?
If you were to build upon this work, what might be the next question that you tackle?
As you can see, the questions are all about the process and reflecting on it. The student might have answers to all of them and it still may not quench your suspicion. And yet, if they can answer many of these questions effectively, then they are likely displaying and demonstrating the things that you are hoping they are able to do.
If they can’t answer these questions effectively or deeply enough, then there might be the opportunity to further explore them reflecting and revising their work with these questions and thoughts in mind. This might deviate from the assignment proper, yet you can also ground it in the fact that the higher learning that is supposed to take place (e.g. the course outcomes/objectives) was not demonstrated clearly through the work itself in conjunction with the conversation.
This can feel like shaky grounding on its own, but this is where you would (gently) lean into your professional expertise as an educator to say that something did not align with the other work the student had done or how the student had previously engaged with the work. You wanted to learn more and allow the student to discuss it to help support the learning process. The conversation demonstrated that some more was needed by the student to better demonstrate their learning.
Again, this might not work every time and it may be tricky to navigate. It requires patience, understanding, the benefit of the doubt, and a willingness to set our egos aside (something that’s always hard because teaching is so personal for many of us). Yet, this is likely to help the student better understand your and your approach to teaching and learning, how they work, and how they figure out their own means of working and learning.
And Yet…
I firmly stand by what I’ve shared here as one of the better ways to navigate concerns about AI plagiarism (or any “academic dishonesty”). Yet, the real challenge is that such approaches require two things from us that are in such short supply.
Attention: It requires us to focus our attention, emotions, and thinking to work through a process of engaging with a student which for may educators is natural and for others, it can be more challenging (especially where you are on the continuum of viewing “academic honesty”). This attention can take a lot of mental and emotional bandwidth to maintain, especially if you finding more than 1-2 students in this spot.
Time: Coupled with attention is just time. These conversations take time. They take time to prepare for, time to schedule, and time with the student as well as time in the follow up. Again, more than 1-2 students in a course, and you’re looking at hours more time. It’s not that this isn’t time well spent—there is more learning that is happening, but it often comes at a cost of something else that does happen or gets pushed back or some other way of showing up for students is diminished.
It comes down to the challenge of “one more thing” that needs to be done, cultivated, followed up on and the like. And I don’t have much advice for navigating that. That’s a larger cultural result of being in a hyper-productive capitalist society that values (questionably-valued) outputs over process; and if you got a solution for that, I’d love to hear it!
The final post in this series will how to actually jumpstart the conversation about generative AI in your classes from the get-go.
AI+Edu=Simplified by Lance Eaton is licensed under Attribution-ShareAlike 4.0 International
Keeping going with this fine series, Lance.
Heard an anecdote. One university manager said they knew a certain staffer used ChatGPT to write something - because the results were good.
You're on to something here, Lance. Your first essay in this series came up in a meeting I was in yesterday. I bragged to someone later that I was one of your early subscribers!
Here's why the series is resonating with me. I think the worst impacts of generative AI have been to push us to surveil our students using monitoring software (powered with AI!) to keep them on the straight and narrow path. Jason Gulya has a good essay up on The AI Edventure on how Coursera's latest announcement is pointing in this direction.
https://higherai.substack.com/p/meet-the-new-face-of-traditional
In my most dystopian moments, I imagine a world of monitoring software combined with predictive algorithms (also powered with AI!) auto-nudging students to engage more with LLMs chatbots in order to demonstrate engagement, forcing us all down a path to AI College from Hell.
Reading your thoughtful approach and guidance to instructors working through these issues is giving me a vision of much more hopeful future.