00:00:00 Welcome to episode fifty of If Teachers Ruled the World. I’m Leah Cleary and we are in the home stretch of our Summer Professional Development series.
We’ve covered a lot of ground this summer: the tech pendulum and why blanket bans miss the point, why ignoring AI in schools is a mistake we’re going to regret, a practical framework for deciding where technology belongs in your lessons, and a wellness episode that I hope gave you permission to close the laptop and see the eagle.
And now, with the school year getting closer, I want to get into something that I think is one of the most practical, immediately useful things you can do before your students walk back through that door. I want to talk about why we need to stop policing AI and what to do instead. I’m going to start this episode with a story that begins with a punctuation mark.
00:02:21 A few months ago, I think it was January, a student looked at something I had written, feedback I had given her, and she said, “Did you write this with AI?” I said no, I wrote it. I mean, it was handwritten. It was on her paper. Then she pointed out my em dash.
Now, for anyone who doesn’t know, an em dash is a punctuation mark. That’s all. It looks like a long hyphen, and it has become, in certain corners of the internet, the symbol of AI generated writing. The GPT dash, some people call it.
If you see an em dash, apparently, you’ve caught a robot. Here’s the thing. I’ve been overusing the em dash for thirty years now. As a matter of fact, a writing teacher in my college told me so. She said I rely way too much on that piece of punctuation.
And I know why I do it. It’s because I grew up reading a whole lot, and the em dash has been used in novels. It’s been used in plays for centuries. Jane Austen, she was an em dash ninja. So was Shakespeare. It’s still used in novels and modern drama. Playwrights have used the em dash since forever to capture the rhythm of how people actually talk. I absorbed it and I never stopped. And it’s such a versatile little piece of punctuation.
So I’m standing there accused of outsourcing my handwritten feedback on some little tiny assignment to a robot based on a piece of punctuation I learned from Shakespeare. So this is one of those crazy ironies, because that same day that I was accused, I was in my car on my way home from work listening to 99% Invisible.
For those of you who don’t know, it’s a really good podcast, I recommend it. So they did an episode about the em dash. They did an episode all about its long history and how it’s come to be erroneously associated with AI. It’s a really good episode, so I’ll link to it in the show notes. Now that’s what 99% Invisible does. They take a topic like a piece of punctuation and they make it just fascinating. So my first reaction was to laugh.
My second reaction was to think about what my student said. And so if my own writing is being flagged as AI, what does that mean for my students? What does that mean for the whole project of catching AI in student work? And I think it means that the project is flawed, fundamentally structurally flawed. And we need to talk about that.
Here’s what we know about AI detection tools. They do not work reliably. Not even close. The research on this is consistent. AI detectors produce significant rates of false positives, flagging human writing as AI generated.
They’re particularly unreliable for non-native English speakers, for students who write in informal or academic registers, and apparently for teachers who read too many plays in college. I also got flagged one time for using the word “mitigate.” Interesting, right? I use that word all the time.
But beyond the tools themselves, the broader tell approach is a problem. The em dash is a tell. Certain sentence structures are tells. Transitional phrases like “not only but also” are tells. And the more teachers rely on these tells, the more students learn to game them, to deliberately write in ways that don’t look like AI, which is not the same thing as actually learning to write.
I mean, you can take an AI generated essay, take out all the em dashes, replace all the “not only, but alsos,” and it might or might not get flagged. The only reliable indicator is knowing your students, knowing their voice, knowing their patterns, knowing their typical mistakes, their level of sophistication. When something doesn’t sound like them, you notice. But that’s a relationship skill, not a detection tool. And it only works if you’ve been reading their writing consistently enough to know what their voice actually sounds like.
Here’s the real problem with spending energy on AI detection. It puts you in an adversarial relationship with your students. You’re the cop. They’re the suspect. And that framing erodes the classroom relationship in ways that outlast any individual assignment.
I still remember being in seventh grade and writing a book report, and the teacher accusing me of copying it from the back of the book. All she had to do was go look at the book and see that I didn’t, or know me and know my writing. I wrote pretty well for a seventh grader, but she didn’t. She just jumped straight to accusing me. I still remember that. And I still have negative thoughts about that teacher.
See, I don’t want to be the AI police, and I don’t think you do either. So let’s talk about what we can do instead.
00:08:06 Before I really get into strategies, I want to spend a minute on the why, because I think some teachers have landed in a place of “if AI can write it, why are we assigning it?” And I understand that feeling. But I think it’s the wrong conclusion.
Writing isn’t just a product. Writing is a cognitive process. When students write, they’re not just recording what they already know. They’re figuring out what they think. The act of putting ideas into sentences forces a level of clarity and precision that reading and listening don’t require. It’s one of the most powerful thinking tools we have.
So we need to keep writing. Research on writing and cognition is consistent. Students who write about what they’re learning retain it better, understand it more deeply, and make connections across ideas more readily than students who just consume content. Writing is not a way to demonstrate learning after the fact. Writing is how the learning happens.
When AI writes for a student, the student skips that process entirely. They get the product without any of the cognitive work. And that’s not just an academic integrity problem, although it is. It’s a learning problem. That’s a student who’s sat in your class and learned nothing, with a polished essay to show for it.
So the goal is not to eliminate AI or to detect AI. The goal is to make sure the thinking still happens, that students are doing the cognitive work, even in a world where a tool exists that can do the surface level product for them. And the way to make sure the thinking happens is to make the thinking visible.
00:10:11 Here’s the mindset shift I want to offer you. If AI can generate the final product, the final product is not where the learning lives. I’m going to say that again, because I think it’s really important: if AI can generate the final product, the final product is not where the learning lives. The learning lives in the process.
So weight the process, make the process visible, make the process gradable. The simplest version of this is the most effective. Have students write in class in front of you with their devices put up. Grab a pen or a pencil. Grab some notebook paper. Do it every day. Write on mini dry erase boards.
For any high stakes writing, or just daily processing, requiring at least a portion of it to happen in class, where you can see the process and have conversations about their thinking, changes the entire dynamic. You’re not policing, you’re teaching, and there’s a real difference there.
For bigger projects, a process journal where students document their brainstorming, their research decisions, their revisions makes the thinking visible in a way that’s hard to fake. And rubrics that weight process more heavily than the final product send students a clear message about what you value. If sixty percent of the grade is on the process and only forty percent is on the final paper, students have a much stronger incentive to actually do their thinking.
This is the approach I’ve found to be the most powerful. It’s about deliberately sequencing when technology enters the process, not banning technology, but deliberately sequencing when it’s used. It looks like this. Students brainstorm on paper first, no devices, just thinking and writing by hand. Then they open their devices to type and develop their ideas. Then peer review.
Then, and only then, do they bring in AI, not to write for them, but to give them specific prompted feedback. What’s unclear in my argument? Where do I need more evidence? What’s the strongest part of this? What’s a good counterargument?
And then students revise based on that feedback, using their own judgment about what to keep and what to change. In that sequence, AI is a feedback tool, not a ghostwriter. The student did the thinking. The student wrote the draft. The student is making the revision decisions. That’s ethical AI use in practice, which is actually what I talked about earlier this summer.
00:13:06 I want to take a minute here. I don’t typically do this, but I want to be upfront that this is a promotion, though I think it’s a very useful one for where we are right now, because I really think teachers need assistance with this.
The school year is right around the corner, and I have three courses that I want you to know about before you walk back into that building. They’re all digital, they’re all self-paced, and they are all going to be very, very helpful for you.
The first one is brand new. It’s called Reimagining Instruction in the Age of AI. It’s a course that ties most directly to everything we’ve been talking about this summer. It’s got four lessons, but the lessons are divided up into micro lessons. I know you don’t have unlimited time, and it walks you through exactly what it looks like to shift your instruction from product to process, design assignments that make thinking visible, and integrate AI as a tool rather than treating it as an enemy.
This is the practical, hands on version of the philosophy I’ve been laying out all summer, and I really believe in it. This course has lots of bonuses that go with it that are going to help you put this into practice in your classroom.
The next one is Smart Classroom Gamification. It’s a course design and classroom management approach that I’ve used for years, and it has been really successful for me, both at the middle and the high school level. The concept isn’t that you’re playing games, it’s that you turn your course into a quest.
Students earn points, they level up, they work towards goals. It is so much fun and it really changes the entire energy of the room. The course builds you up from the ground, gives you Google Sheets that auto tally points, content templates, and AI prompts to help you design your quest that integrates seamlessly into your course. If you want something that genuinely re-energizes both you and your students heading into a new year, this, my friends, is worth your time.
We’ve also got the Complete Unit Planning System, and that is for the Sunday Dread that every teacher feels. You know the one. It’s the end of the weekend, the week is coming, and the plans aren’t done.
This course gives you templates that do the planning thinking for you. You build out your units once, standards aligned, with activities linked and ready, and then your weekly planning becomes plug and play. Heading into a new school year is exactly the right time to set this up.
All three courses are available right now at leahcleary.com/courses-for-teachers. And through July 31st, you can get 15% off any of them with the code SUMMERPD26. That’s S-U-M-M-E-R-P-D-26. It expires at midnight on July 31, 2026. So if this is hitting your ears close to that date, don’t wait. Links are in the show notes.
00:16:33 Let me bring you back around to where we started. My em dash. My thirty year habit inherited from playwrights and novelists. The em dash, I think, is a useful tool. I hate that it’s been vilified because AI overuses it.
And that’s a little absurd, because the problem was never the em dash. The problem is thoughtless use. Just like the problem with AI in your classroom was never AI itself. The problem is thoughtless use. Students using it to skip the thinking. Teachers using detection tools to play police instead of doing the harder, more important work of teaching the process.
We can do better than that. And heading into this new school year, with everything you’ve been thinking about this summer, I know you want to. Stop being the AI police. Let’s just be the teacher. Make the process visible, make the thinking count. And maybe, and this is a plea from me, save the em dash while you’re at it.
00:17:46 Season two is right around the corner. I’ll see you really, really soon. But before that, we have two more summer episodes. Next week I’m coming to you with some teachers’ opinions on technology in the classroom.
And the week after that we’re going to have a conversation with Adam Sparks. He is in the edtech world and he is the inventor of one of my favorite tools, Short Answer. And he has some really important things to say about edtech for us. So I think you’re definitely going to want to hear that one.
All right, I’ll see you very soon. And don’t forget to follow me wherever you get your podcasts. Leave me a review if you liked this episode, and I will see you next week.
