Are students really cheating their way through college?
No.
I mentioned in my last post that we should expect AI research to be messy. The same goes for when we see a hyperbolic clickbaity article about how all students are cheating. It’s just not that simple. Sure, some students are cheating—but they always did. Some students are so afraid of the specter of cheating that they record themselves working or they avoid even touching AI (to the extent that they can!). The majority of students are somewhere in-between, experimenting with AI tools and justifying what they do with reasons both legit and sus.
I know this because I’ve talked to my own students, surveyed and run focus groups of students at my university, and read studies of what students are doing with AI more generally. The way we can understand how students are actually using AI is to 1) talk to them; 2) look at data; 3) collect data ourselves.
Students want to talk to their instructors about AI. They talk to their peers about it, but they’re navigating such a complex ethical space that it’s hard for them to do it with peers only.
To better understand what students are doing with AI, I looked at data. I’ll share a link below to an analysis I did for the Norton Substack, AI & How We Teach Writing, in which I analyzed 10 different data sources and reports on AI use among students. It’s messy. Go figure!
Finally, nothing beats local data. Results from a survey I’ve given to Pitt students since Dec 2022 have been profoundly useful to my own teaching, as well as for crafting policies and responses at Pitt more broadly. If you’d like to survey your own students, you’re welcome to go ahead and adapt my survey.
Stories are great and all, but let’s not fall into the trap of believing journalistic anecdata instead of listening to our own students.
I love the term 'anecdata' - captures a real problematic feature of intellectual life that otherwise escapes notice or passes as fact, though a good anecdote always stands on its own!
what is the function of a writing assignment in English or a problem in mathematics. In math, at one time, students provided hand-calculated answers and today they are required to have programable calculators for exams. For English it might be to show understanding of a story. Most of these are to complete an assignment. Once completed, they will do a cognitive offload and pick up the next assignment. In the new world where AI tools replace old methods, the students demonstrate solutions within that new knowledge matrix demonstrating process learning and applications to problem solving. Unless the lessons learned are carried both in content and process throughout the education, those students who step outside the box and can transcend the current system will, as they advance, leave their cohorts behind and the teachers will find that their future will also stagnate