I set the timer and focus on my new goal: to grade a paper in twenty minutes. Everything starts smoothly; checkmark after checkmark, praise after praise. A quick glance at the rest of the paper reveals more of the same error-free, polished style. Hmm. I pull up the student’s writing diagnostic, previous assignments, and emails. All are riddled with errors and awkward phrasing. The plagiarism report, though, comes up clean.
Now what? Do I press on and gamely mark the rest of the paper, knowing in my gut that the student hasn’t written it? Or do I double-down and find proof of academic dishonesty? The first option is certainly the easiest. I can whiz through, click the top boxes on the rubric and move on to the next essay. No fuss, no muss. Nevertheless, it feels disingenuous. I am not doing the student any favours by playing along. And it seems a disservice to the others who have worked hard on their papers, so I go with plan B.
First, I check the file properties to confirm that the author is indeed the student. I then copy and paste phrases from her work into Google. This doesn’t yield anything, but when I plug the passages into an AI detector they flash red, signaling a high probability of AI content. Like a gumshoe detective, I start a folder for the “case” and fill it with screenshots and notes. Next, I email the student to set up a meeting. Total time on my first paper: one hour and ten minutes with twenty-two more yet to grade. Gulp.
I carry on with the rest of the essays and curtail the suggestions, underlines, and highlights, but I still don’t make my lofty “20-minute-per-paper” target. Thirty-five to forty minutes seems to be the average until I come across another gold-star piece of writing worthy of a graduate student. Then another and another. What is going on? A couple of days later, when I finally finish the grading, I feel like a train wreck. The back and forth with the students is exhausting. I don’t want to accuse them without hard proof. AI detection software is still new and not always reliable. Also, where do I draw the line? Are Grammarly and QuillBot okay for certain assessments or should they not be used at all? To add to the AI frustration, there are countless paper mills or “contract cheating” services in Canada. In a recent Globe and Mail article, Joe Friesen writes, “Allegations of academic misconduct – meaning unauthorized help, cheating, plagiarism and other acts of dishonesty or misrepresentation – have risen sharply at many schools.”
One by one, I meet with my students on TEAMS. Psychologically, I can’t add these hours to my grading total because it would make me insane. I show them my evidence, whether it be the red-hot results from an AI detector or an unrecognizable author’s name. Or, the most reliable, a side-by-side screen view of their original writing against their newly-minted work. Usually, they fess up quickly with a host of excuses and apologies. And most of the time, I give them a second chance.
Curious about what is driving the increase in academic dishonesty, I chat informally with the class (which is in person) about their motivations. A common theme quickly emerges among international students. The high fees and increasing cost-of-living expenses force them to get a job, which takes away time from their studies. In addition, many struggle with low confidence in their English abilities and a fear of failing the course. Add to this the homesickness and pressures from home, and the terrain is ripe for cheating.
What is the answer? My mind scrambles over the possibilities. Could we develop an ungraded, merit-based system, perhaps? One where students, individually and in groups, progress through a series of learning activities earning a point for each unit that they fully complete in class? In elementary school, we had a program where we chose books from designated boxes that matched our reading level, and once we met the target we graduated to the next box and so on. I remember the drive I had to reach that next box and the feeling of pride when I got there. Sure, there are problems with this scenario, such as the frustration of those who remain in one box while their peers move to the next. Nevertheless, we teachers would be there to provide feedback, motivation, and support. We’d also have more time and energy to devote to the students, having freed ourselves of hours of marking.
All I know is that I do not want to be a zealous, “I caught you red-handed” type of teacher. I want to spend less time pouring over essays and hunting for cheaters and more time generating creative lessons. I want to, well, do what I am good at and what I am trained to do.
Jennifer Hutchison
I’m Jennifer Hutchison and I teach EAP and communications at George Brown College in Toronto. I have also taught courses in sociolinguistics in the English Foundation Program at Toronto Metropolitan University. In my spare time, I write short stories, read, exercise, and bake (the last two are codependent). Teaching English is my passion. I am curious about the world around me and feel fortunate to have that world brought to me every day in the classroom. Nevertheless, I took a circuitous route to discover this passion. After my undergraduate degree in French and translation, I worked as a translator and then veered off into writing and editing, which I did from home while I raised my children (four of them!). In none of these positions (except, possibly, childrearing) was I helping anybody, so I returned to school, launched my ESL career, and have never looked back. I look forward to working with you and sharing experiences and strategies on the Blog!
Educators *everywhere* are with you here. AI has so much potential, but man-oh-man, the temptation to unplug one’s brain and substitute for whatever ChatGPT spits out is unreal. You nailed it – our students face so much pressure and have so much to lose, and add that to the fact that they lack confidence in their own abilities, that yes, they will yield to AI.
I love your line about being a “gumshoe detective”; I did an activity once with my class called “If You Can Find It, So Can I” and showed them how a teacher (or employer) might track down the original source if they suspect it’s not the individual submitting it. It didn’t seem to help.
Students have always been able to find a way to submit work that isn’t entirely their own; wether it be AI-generated work, or a paid human “tutor”, this is a problem that has always existed.
I love that AI can support, guide, and suggest revisions to a student’s work. I don’t love that AI can do the work entirely. So what do we do now? That’s a question that’s loudly ringing in the education-sphere (join ANY AI/ChatGPT support group and it’s by far the MOST common query).
The answer is what you touched on above – ungraded, merit-based systems, or project-based learning that encourages peer feedback and self-reflections on target learning objectives. Alternatively, pen and paper assessments might see a surge (until some genius figures out how to upload ChatGPT directly to our frontal cortex…).
Plug AI/ChatGPT into the SAMR model – maybe we shouldn’t be going for a pure substitution here. Education and learning is being REDEFINED in our era. It’s exciting and a tad scary, but we’re helping no one by chasing our tails and trying to impose 20th Century education standards on what’s going on in today’s classroom.
Nicely said, Jan!
I am with you.
I love your line about using pen and paper “until some genius figures out how to upload ChatGPT directly to our frontal cortex…” Who knows, maybe that comes next! LOL