00:00:00 – Leah Cleary
Before we get into today’s episode, grab this. I’ve got a free four move framework for AI proofing any assignment so students do the real thinking on their own first, and then you teach them how to use AI to augment what they’ve already built, not replace it.
Go get it right now at leahcleary.com/ai-proof-framework. While you’re there, you’ll also see my complete course, Reimagining Instruction in the Age of AI. It gives you a framework for teaching ethical AI use, exactly what states are starting to require as part of AI literacy.
Plus it walks you through AI proofing your own assignments step by step with AI prompts to help you actually do it. So go grab that now, because we are about to talk about why this matters more than ever, with AI policy changes on the horizon at the federal level and in states across the country.
You are not going to want to miss this. So go to leahcleary.com/ai-proof-framework. And I’ll see you in just a minute in the episode.
00:01:53 – Leah Cleary
It’s time for your monthly teacher reality check, where we translate education policy jargon into what it actually means for those of us in the classroom.
So I just started at a brand new school, in a brand new grade level, teaching a brand new subject. And one of the things I’ve noticed fast is that my students don’t really understand the difference between using AI ethically and using it to just make their work disappear.
You see, I built a Gemini Gem for my students to use at home. Something to walk them through some pretty complex study guide questions, not to give them the answers, but to help them think through the answers.
And when I introduced it for the first unit, more than one student looked at me and basically said, oh, wait, you’re letting us cheat. I had to stop and say, no, I’m not letting you cheat.
I built this Gem and it’s the only chatbot you’re allowed to use. Honor code. It is not going to give you the answer. It’s going to walk you through dissecting the prompt and arriving at your own conclusion, the same way I would if you raised your hand in class.
This just means you have the same help available to you when you’re not in class. Well, maybe not the same, but something. And to me, that is actually the real upside of AI done right. It levels the playing field.
Not every kid has a parent who can sit down and walk them through a hard assignment, or the money for a tutor. A tool like this means the kid whose parents are working two jobs gets the same kind of support as the kid whose parents can afford that extra help. That’s not cheating, my friends. That’s access.
But my students’ reaction told me something very important. Nobody has drawn the ethical line for these kids. They don’t know there’s a difference between using AI to think and using AI instead of thinking.
And on top of that, a lot of them don’t seem to know, or they don’t want to know, that AI is frequently wrong. Frequently and confidently.
I caught this myself before I moved a lot of my writing back to pen and paper in class. I had a student in AP World History turn in an essay that did not sound like them at all, and it had a glaring historical error.
See, they used the Song Dynasty, which fell to the Mongols in the twelve hundreds, as an example of something happening in China in the fifteen hundreds. Off by about three centuries and a different dynasty entirely.
That’s the kind of mistake that happens when a student hands the thinking over to a chatbot and doesn’t look closely enough to catch it.
Meanwhile, I do use AI myself. I use it to brainstorm, to think out loud when I’m stuck, to hand a rough draft to and ask, where is my argument weak? What am I missing? It’s never once done my thinking for me. It’s a second set of eyes, not a replacement for the first set.
That gap, between how I use it and how my students assumed it was meant to be used, is exactly what state legislatures are starting to catch up to.
00:05:36 – Leah Cleary
This year alone, at least one hundred and thirty four AI and education bills have been introduced across thirty one states. Thirty one states now have official AI guidance out from their departments of education.
At the federal level, in April of twenty twenty five, the president signed an executive order called Advancing Artificial Intelligence Education for American Youth. It created a White House AI education task force and directed the Education Secretary to prioritize federal grants for teacher AI training.
Now, separate from that, and I want to be really clear, these are two different things. There’s PISA. PISA stands for the Programme for International Student Assessment. It’s the exam that already ranks fifteen year olds across the world in reading, math and science every three years.
Starting in twenty twenty nine, PISA is adding a media and AI literacy category for the first time ever. That’s not a US policy. It’s run by the OECD, an international body, not the White House.
But it means how American students measure up on AI literacy is about to become part of a global scoreboard. The same scoreboard where we’re now sitting around twenty eighth in math.
At the state level, Idaho has built a full AI framework covering privacy, procurement, transparency, academic integrity, literacy standards, and teacher training.
Maryland’s AI Ready Schools Act requires districts to adopt a policy, name an AI coordinator, and, this is the important part, pays for statewide teacher training. Virginia is funding and evaluating AI pilot programs in schools.
Mississippi and Alabama are tying AI instruction to graduation requirements through computer science credit. California is protecting student data, prohibiting it from being used to train AI models.
Oklahoma wants parents notified when AI tools are used with their kids, with an opt out option. Ohio is leaving it to individual districts to write their own policy, and West Virginia is the one to watch, guys. Their bill actually ties district funding to whether districts adopt a compliant AI policy.
00:08:19 – Leah Cleary
So here’s your reality check. Maryland and Virginia are the exception. Most of these laws hand down the mandate and stop there. Teach this, train on this, build a policy around that, and find the time and the budget yourselves, using money that was already stretched before anyone said the words artificial intelligence.
Now, there’s one detail out of the Idaho legislation that I really can’t stop thinking about. Written into the actual bill, in legal language, is a line that says AI shall not replace human teachers.
My first reaction was, wait, who thought that was on the table? But we’ve all actually seen this before. Back between about twenty fourteen and twenty eighteen, when schools were rolling out one to one devices, I remember teachers saying almost exactly this. This software is going to replace us.
And then the pandemic hit, and kids really were placed in front of software at scale, with minimal live instruction. And we all watched it fail, not because the software was bad, but because a program can’t notice a kid is having a rough week and adjust on the spot. A teacher can. That fear didn’t survive contact with reality.
So when a legislature feels the need to write that AI won’t replace teachers into an actual bill, that tells us the fear is still loud enough that lawmakers felt they had to say it again. But I don’t think the law is what actually settles that fear. Using the tool does.
I use AI and it has never once made me less necessary in my own classroom. If anything, the teachers who get their hands on it are usually the ones who stop being afraid of it, at least for now.
So here’s where this leaves us. This is coming, whether your district has said a word about it or not. Some states are funding the transition. Most aren’t. And if nobody at your school has handed you a policy or training yet, that’s not you falling behind. That’s a predictable gap in how this is being rolled out, and you’re allowed to name it exactly as that. It’s a gap.
00:10:58 – Leah Cleary
So the question is, what do you actually do Monday, regardless of what your state or district has or hasn’t announced yet? First, model your own AI use out loud, transparently.
I’ve done this before with freshmen using Google Slides, where you can show version history. Here’s the slide I built from scratch. Here’s the version after I asked AI to help me tighten it up. Same content, mine, just polished and more graphically sound. Showing them the receipts teaches the ethical line better than any rule on a poster.
Second, and this is the piece I built out fully in my course, Reimagining Instruction in the Age of AI, there’s a simple test you can run on any assignment. Could a student finish this with AI alone, start to finish, with no thinking of their own required? If the answer is yes, here are four moves that put the student back into the work.
Move one. Make the process visible. Before a single device comes out, students start a research and decision log. What sources they used, what they rejected, why they rejected it. It gets turned in with the final product, not as an afterthought, as part of the grade.
Move two. The AI layer. AI only enters after students have done their own initial thinking, never before. It checks for gaps and accuracy. Students document what it told them and what they did with that feedback. AI can inform the product, it cannot write it. And we all know that sometimes it will be wrong about fact checking.
Move three. One human layer per product. Every assignment gets one required addition AI cannot supply. A real perspective, a lived detail, an original judgement call. Specific beats vague. Include a quote from an eyewitness you name and describe. See, that works. Be creative doesn’t.
Move four. The oral component. During presentations, every student answers one process question. Cold. No prep. No notes. Not a trivia check, a process check. A student who actually did the work should be able to easily answer it.
I built a free one page breakdown of these four moves, and I want you to have it. Go grab it at leahcleary.com/ai-proof-framework. It’s free and it walks you through all four moves in more detail than I can give you here.
And if you like what you see there, that’s also where you can learn about the full course, which walks you through not just AI proofing assignments, but teaching students to use AI ethically in the first place, in ways that are in line with a lot of what these new state mandates are now requiring anyway. Go grab it. Don’t put it off.
00:14:12 – Leah Cleary
So if I ruled the world, this one’s easy. There would actually be funding attached to these mandates. Because what we’re looking at right now is, once again, states and the federal government asking teachers to do more, without giving us the time, the training, or the money to do it well.
Until next time.
