On Attribution and AI
How do we tackle academic integrity and attribution in an AI-forward classroom?
TL;DR (too long; didn’t read) - Acknowledgments (instead of citations) can be a much more sustainable approach towards setting expectations on AI attribution. Here’s a guide to learn more on how to implement it.
I first started thinking about this topic when I decided to openly use AI tools in my classes. I had been using AI tools myself for some time and with the start of a new semester, and the freedom I had in my elective courses, I thought it’d be a good idea to introduce my students to AI tools and integrate them into the work we were doing in class. (For the sake of keeping the story simple, I’ll be talking about this topic in the context of my Programming & Electronics class.)
By this point in time, ChatGPT had been released for over two years. Students were using it, teachers were using it. We were all using AI. But was there a structured approach? A perfectly laid out way of integrating it into the classroom? No, of course not! This stuff is hard and AI moves fast! What was clear was that in the context of my course, I wanted to offer my student opportunities to co-create and collaborate with AI. The big question though was - what do I do about attribution in a situation like this? Like with many things AI, it forced me to face a new scenario that I hadn’t come across before.
My first thought was, “I’ll just ask students to cite AI”. After all, most of the major academic citation styles had already published guides for citing AI. But then I realized, with the type of work that I was asking students to do (brainstorming ideas, getting feedback on prototypes, co-writing code, etc) they weren’t necessarily taking exact content from the AI, so citation wasn’t exactly appropriate. It was blurred and there was no clear black and white on what was AI generated and what was student-generated. The traditional citation model did not work for us.
After lots of researching and exploring what other academics, writers and educators around the world were doing, I started to notice that the use of brief notes was quite common when it came to AI. In footnotes, in intros, in closing notes, etc. Reading lib guides from major university libraries around the world, I came across the word acknowledgment quite a lot. Is it new? No. Acknowledgments have been a part of academic attribution for a long time. But in the case of AI attribution, it seemed to be a nascent concept. How often do we really use the exact output of AI, as is, in and out, raw, without re-prompting? How many times have you entered an initial prompt, gotten a reply and thought, “hmm, let me try again” and then proceeded to add a lot more context to your prompt to get a much higher quality result? I would bet, a lot.
So eventually I started to think of attribution in these two broad categories - acknowledgment and citation - and laid it out as such with my students. “If you’re ever taking the exact output of an AI, I expect a formal citation. If you’re using AI throughout the process as a tool, I expect you to give credit to the AI in the form of an acknowledgment.” I realized shortly thereafter, that I was not the only one in this situation. My colleagues across the high school were having similar challenges and so I was asked to give a short workshop with teachers on how to deal with AI attribution in the classroom.
The resource I’m sharing here is an updated and further refined version of that workshop. It gives actionable examples and provides thoughtful guidelines on how we should be asking students to attribute AI use in their work. Is it perfect? Absolutely not (in fact, I would love to hear your feedback!). But it is something you could implement in your classroom immediately and start being open and transparent around AI use while maintaining a high standard of academic integrity in the classroom…even in the age of AI. This whole topic might prompt thoughts around AI cheating, but I won’t get into that now - it’s beyond the scope of this current monologue - but since I’ve mentioned it, just a reminder…AI checkers do not work ; )
Now back to this resource. It’s meant to be used as you see fit. Whether it’s for your own use, to learn at your own pace, or even to use in the classroom and show your students, be my guest! The ideas are not entirely my own (I’d encourage you to check out the sources) - they are simply an interpretation of the varied resources I’ve found for a potential approach we could take on attribution and AI. Enjoy and please reach out if you have any questions or feedback!
Note: No AI was used in the writing of this post. If I made grammatical errors, awkward word choice or you find other clues of less than perfect writing - forgive me!
Know someone who’d find this useful? Forward the email or share this post!
Forwarded or shared with you? Click the subscribe so you don’t miss the next one.
Find me across other social media platforms here :)