AI in Education: Shifting Perspectives for Better Learning

When I was still in the classroom teaching French, my students had a predictable habit: they tried to use Google Translate to “cheat” on their writing assignments. I’d collect beautiful, complex sentences from students who could barely remember the word for “apple.” It was obvious. But rather than get angry, I remembered something my mentor Deb Blaz, truly the best teacher I’ve ever known, once told me:

“Don’t fight the tool. Teach them how to think with it.”

So instead of banning Google Translate or turning assignments into surveillance exercises, she created an activity that changed my teaching and shaped the way I think about AI in today’s classrooms. I used this activity with my students annually.

We took a trip to the computer lab (yep, that long ago) and gave students ten sentences in English filled with idioms and common conversational phrasing. Their job was to use Google Translate to convert them into French. Then, take those French translations and run them back into English.

The results? Hilarious!

“He kicked the bucket” came back as a literal image of violence. “I’m pulling your leg” returned something about limb removal.

Students were howling. But more importantly, they got it. They saw, with no lecture needed, that translation tools don’t always understand meaning and their work couldn’t be faked that easily.

Then we leveled up. Students wrote their own paragraph, translated it to French, then to another language of their choice, and then back to English. What came out? Total nonsense. But through the laughter, something serious was happening:

They were learning how tools work, and more importantly, how to work with them.

Fast-forward to today, and it’s not Google Translate, it’s ChatGPT. It’s AI detectors and Turnitin reports and hallway hand-wringing over what’s real and what’s machine-generated. But the root issue is the same:

–AI isn’t the threat. Misuse is.
-Cheating isn’t new. The tools are.
-We can’t just ban; we have to teach.

Students today need the same type of structured sandbox that Deb gave me. A chance to explore the tools for themselves, to hit the limitations, and to unpack the results together.

That’s how we move from fear to fluency.

So here’s what I’d suggest for any educator navigating the AI landscape:

Teach the Tool Before You Test the Task

Create low-stakes space for students to use AI. Let them prompt it, break it, laugh at it, learn from it. Then teach them how to evaluate its output—not just accept it. Students are more savvy than we give them credit for when we let them be.

Shift the Assignment, Not Just the Penalties

If an AI can do your assignment better than your student can, the assignment needs work. Focus on voice, context, connection, and critical thinking. Ask students to reflect on what they learned, why it matters, or how they got there—not just what they produced.

Make Academic Integrity a Dialogue, Not a Disclaimer

Talk to students about why integrity matters. Show them how tools can support their learning rather than short-circuit it. Build relationships so that when they’re tempted to cheat, they come to you instead.

I don’t believe students don’t want to cheat. They want to succeed.
They just need help learning how to do that in a world full of shortcuts.

And let’s be honest: we use shortcuts all the time. We crowdsource ideas in group chats, ask Gemini to polish our emails, run slide decks through AI design tools, and Google the answers we’ve forgotten. We call it collaboration. We call it efficiency.

But when students do the same? We call it cheating.

If we want to raise critical thinkers and ethical leaders, we have to stop moving the goalposts. Let’s stop punishing students for using tools the way we do and start teaching them to use them with purpose, curiosity, and integrity.

They might just surprise you.
And if not? Well, that’s not a scandal; it’s a teachable moment.
Fire up the virtual “computer lab,” grab some popcorn, and let them watch AI fumble their assignment in real time. Nothing builds buy-in like a badly translated metaphor and a robot-generated essay that sounds like it’s auditioning for a soap opera.

And let’s not pretend the adults have it all figured out. Just check your inbox or scroll your social media feed, and you’ll find plenty of copy-pasted AI content, half-baked captions, and emails that read like ChatGPT wrote them at 2 a.m.

So let’s skip the shame and lean into the learning. Because laughter + reflection? That’s the sweet spot.