The Learning Curve #1: Most colleges are policing AI. The better job is teaching it.
Welcome to the first issue of The Learning Curve. If you subscribed when this was The Planning Window, you are in the right place: same inbox, same promise, a slightly wider beat. Every other Thursday I will send one thing worth your time from inside the work, rotating across teaching with AI, the AI tools and models actually worth your attention, digital accessibility, and spatial learning. One field note, one move you can make in five minutes, one link. No hype, no hard sell.
What I'm seeing
Almost every AI conversation I have with faculty starts in the same place: how do we catch it. The detector got switched on, the syllabus grew a new paragraph of warnings, and everyone is a little more tired than before.
Here is what those conversations keep circling back to. The catching does not work, and it costs more than it returns. The tools are unreliable enough that major universities have turned them off rather than defend them in a grade appeal. And every hour spent building a case against a student is an hour not spent teaching that student to use these tools well. Only one of those is a fight you can win.
So the shift I keep coming back to, and the reason this newsletter exists, is easy to say and harder to do. Most colleges are spending their energy policing AI. The better, more durable work is teaching it. You do not have to AI-proof your course. You have to make your assignments worth doing in a world where the tool exists.
Try this (five minutes)
Add an AI Reflection Statement to your next assignment. Students submit, alongside the work, a short note on how they used AI and what they changed about its output. A prompt you can paste in today:
"In 4 to 6 sentences: Did you use AI on this? For what? Show me one thing it gave you that you kept, and one thing you rejected or rewrote, and why."
It costs you a sentence in the instructions and a minute of reading. A student who leaned on AI thoughtlessly cannot fake a good reflection, and a student who used it well shows you judgment you actually want to see. It turns "did they cheat" into "show me how you thought," which is the better question anyway.
Worth your time
The Stanford study that should reframe the whole detection debate: Liang and colleagues found AI detectors flagged about 61% of TOEFL essays by non-native English writers as AI, and almost none of the essays by native writers (GPT detectors are biased against non-native English writers, arXiv:2304.02819). If your campus still runs detection, that is the equity problem sitting underneath it. Free, no signup.
P.S. If you read something here you disagree with, reply and tell me. I would rather be corrected by someone in the role than be polished and wrong. And if a colleague owns this work, forward it their way.
P.P.S. Housekeeping: The Planning Window is now The Learning Curve. Same list, same promise, a wider beat. Nothing you need to do.
— Brooks Winchell. 21 years inside higher education, four of those at Quinsigamond Community College.