By Gavin McKiernan
Last September, my biggest worry was whether I would have friends in each of my classes.
This year, it’s figuring out whether typing a prompt into ChatGPT before submitting an essay counts as cheating.
And I don’t think I’m alone in this confusion.
Nobody seems to agree on what “using AI” even means. For some, it’s a tool to explain a concept differently or help organize ideas. For others, it could mean generating a large portion of an assignment. Those are drastically different things, yet they all are lumped under the same phrase: “using AI” – which is why this feels so confusing for everyone.
Students say the line keeps moving.
Ask five students how they use AI, and you’ll likely get five different answers.
AI tools can now do practically everything, from brainstorming to writing full essays, so the old idea of doing your “own work” doesn’t have a clear definition anymore.
“I’ll use it to explain a concept I didn’t understand in class, but I would never have it write an essay for me,” said rising senior Luke Davies. “To me, that still feels like cheating, even if my teacher didn’t directly tell us that.”
Others take a different approach. “Honestly, I use AI tools if I’m struggling to generate ideas,” said rising senior Patrick Doyle. “I’ll still write the essay myself, but sometimes I’ll ask it to help me organize my ideas clearly.”
I understand where both students are coming from, and I don’t think either is wrong.
What I noticed talking to people is that most of us have created our own personal rulebook for what counts as too much help. But nobody has actually compared their “rules” to those of others so we all assume everyone else sees it the same way we do. That’s probably where a lot of the anxiety and confusion comes from.
The problem is not that students don’t care about doing the right thing. It’s that the “right thing” hasn’t been clearly defined yet.
Teachers are adjusting too.
Rather than trying to detect AI use after the fact, Peter Thomas, an AP macro economics teacher at Rye High School, has changed how he teaches.
“Instead of asking students to simply produce an answer, I increasingly design lessons that require students to explain their thinking, defend their conclusions, and make connections,” he said. “AI can generate information, but understanding is still something students have to demonstrate.”
To Thomas, the goal isn’t monitoring whether students use AI, but redirecting it.
“AI should enhance learning, not replace it,” he said. “If it helps you brainstorm, organize, or better understand a concept, it’s a useful tool. If it’s doing the thinking you’re supposed to be doing, it’s no longer helping you learn, it’s impeding your learning.”
Rather than trying to catch students using AI, he talks to them.
“I never look at a student’s work specifically questioning if it was AI-generated,” he said. “I prefer to have conversations with them about the topic. A simple discussion about their work usually reveals whether they truly understand it. My goal has always been to encourage thinking and learning, not police technology.”
The problem isn’t really AI – it’s a communication problem that AI happened to expose. Schools have always had unwritten rules about what counts as your own work, and for a long time, those rules didn’t need to be spelled out. AI has changed that overnight, and a lot of schools don’t know how to deal with it.
A lot of students I talked to aren’t trying to get away with cheating; they are unsure where the line is, and that uncertainty is stressing them out more than the actual temptation to cheat.
That would suggest that the fix isn’t stronger detection software. The fix is schools being upfront, specific, and honest about what’s expected, instead of leaving students to guess and hope they didn’t break any rules.
