EdHeroes

What if the danger of AI in learning isn’t that it gives wrong answers — but that it does the hard part for us?

That’s the idea behind Barbara Oakley’s session at the EdHeroes Global Forum: Grokking: What AI Just Taught Us About How We Learn. In AI, “grokking” is the moment a system stops memorizing and suddenly understands. Our brains may work the same way — but only after real effort. So what happens to that moment when AI does the effort for us?

Barbara is a Distinguished Professor of Engineering at Oakland University and the creator of Learning How to Learn, a course taken by millions.

September 24. Online. Register via the link in our bio 💙

Last year’s Forum ran fully online, which is the whole point: a teacher in Nairobi and a policymaker in Manila could join the same session without a flight or a visa. That’s how you get 70+ countries in one audience.

This year runs the same way — online, open to anyone — but the focus has sharpened. We’re looking at what changes when AI becomes the infrastructure underneath education, not just a tool on top of it: how students get assessed, who gets access, how systems actually adapt.

Three tracks, global speakers, and the same open door. EdHeroes Global Forum 2026 — AI and the Architecture of Future Learning — September 24.

Register via the link in our bio💙

Education is entering a new era — and not quietly.

AI has stopped being a tool we pick up and put down. It’s becoming infrastructure: the layer beneath how students learn, how teachers teach, how systems decide who gets what. That shift changes the questions we have to ask, the partnerships we need to build, and the skills we need to grow in ourselves and our learners.

The EdHeroes Global Forum is where that conversation happens — across three tracks, with global voices working at the intersection of AI and education. One day to step out of the daily grind, see the bigger picture, and connect with the people building what comes next.

It’s for educators, academic leaders, policymakers, EdTech founders, researchers, and the students and changemakers who’ll inherit whatever we build now.

Register now via the link in our bio 💙

September 24, 2026 · Online · Free to attend

This is the post that would’ve saved most teachers from quitting AI in their first week.

The pattern’s always the same. Someone tries it, gets a generic or wrong-level answer, decides the tool’s overhyped, closes the tab. Honestly fair — first answers usually are off. The bit no one mentions is that they’re not supposed to be the right one. They’re the starting point.

Real AI use looks like a short back-and-forth. You ask, you read it, you tell it what’s wrong, it tries again. Three exchanges almost always beat one perfectly-crafted prompt. And the corrections are short — usually one sentence. Five of them are in this post, ready to copy.

The shift is small but does most of the work: you’re not the user. You’re the editor.

Coming up next: a beginner’s map of which AI tool to open for which job — Claude, ChatGPT, MagicSchool, Diffit, Curipod, and when each one actually makes sense.

Which fix do you wish you’d known earlier? Or — what does AI keep getting wrong for you? Tell us 💙

The gap between “AI is useless” and “AI just saved me an hour” is usually about four seconds of typing.

In Part 2 we set up your context. Now: how to ask. The pattern’s not complicated — tell AI who to be, what you want, a bit about your class, and how you’d like the answer back. You don’t need all four every time. But every piece you add sharpens what comes back. The bad-vs-good example in the slides shows it pretty plainly.

The five prompts in here are built to be stolen. Copy them, swap in your grade and topic, keep the ones that work in a note on your phone. That’s a prompt bank, and it grows every week. Two or three usually become daily habits within a month.

Reminder from earlier in the series: whatever AI gives you is a first draft. These get you most of the way. You do the last bit.

Next post: when AI gives you something wrong, boring, or way off — the one-sentence fixes that turn it around. Honestly, that’s the part that separates people who stick with AI from people who quit.

Which prompt are you trying first? And what would you want a prompt for? Drop it below 💙

If you tried AI once, got a bland answer, and decided it wasn’t for you — this is the post that changes your mind.

Here’s what no one tells beginners: the tool isn’t the problem, and neither are you. AI starts every conversation knowing nothing about your classroom, so a vague question can only get a vague answer. Good setup fixes that for good — and it has three layers, each saving you more time than the last.

First, context: tell it who you teach and how you want answers, and everything improves instantly. Second, settings: in Claude and ChatGPT you can save that context once so you never retype it. Third, Projects: a feature in both tools that lets you keep everything on one theme — your Grade 4 science work, say — in a single organized space with its own instructions and files, so every chat there already knows the background.

That last one is the quiet time-saver most beginners never hear about. Instead of one endless messy chat history, you get tidy spaces per subject or task, each pre-loaded with what it needs to know. Five minutes of setup now; hours saved over a term.

What would your first project be? Grade-level planning, parent comms, rubrics? Tell us — we’ll show real examples in an upcoming part. 👇

Most “AI for teachers” advice skips the only part that matters for beginners: the first step.

So before any tools, prompts, or settings — here’s the one thing to understand. AI isn’t a search engine you query. It’s more like a fast, eager assistant you brief. The quality of what you get back depends almost entirely on how clearly you ask. That’s good news, because clear instructions are something teachers are already excellent at.

You don’t need an account full of settings or a single technical term to begin. You need 60 seconds and one honest question about your actual day.

The teachers who get the most out of AI aren’t the most “techy.” They’re the ones who treat it like a conversation — ask, react, ask again — instead of expecting one perfect answer on the first try. We’ll get to all of that in this series.

For now: open it, type one thing, see what happens. You genuinely cannot break it.

Brand new to this? Tell us what’s held you back — we’ll cover it in an upcoming part. 👇

You’re already exhausted. The last thing you need is another mandatory training session that eats your weekend.

Traditional teacher professional development wasn’t designed for real life. It assumes you have time, energy, and the mental space to sit through hours of content that may or may not apply to your classroom.

But professional development doesn’t have to feel like a second job.

Here’s what actually works:

Micro-learning: 5-10 minute modules you can complete between classes. Platforms like Edpuzzle, Modern Classrooms, and Code.org offer free, self-paced training on topics you actually choose. No full-day workshops. No mandatory attendance. Just short, focused learning when you have time.

Peer communities: Sometimes the best PD is just hearing what’s working in someone else’s classroom. Platforms like Edutopia and TeachThought share real teaching strategies from real teachers.

Follow educator hashtags like . Learn from people doing the work, not consultants explaining it.

Teachers deserve professional development that respects their time, expertise, and reality. You shouldn’t have to sacrifice your weekends to grow professionally.

Save this post. Try one resource. See if it fits your life.

What type of PD actually works for you? Let us know in the comments. 👇

Everyone’s asking if AI will replace teachers.

New OECD research found something more important: AI can boost test scores without improving actual learning.

The difference? Whether a teacher is guiding the process.
When students use generative AI without pedagogical support — no teaching principles, no structure, no guidance — they perform better on assignments. Their grades go up.

But they’re not learning more. They’re outsourcing the work, not building the skill.

When teachers guide AI use — setting clear goals, framing how to use it, connecting it to learning principles — students perform better AND learn more. The AI becomes a tool for thinking, not a shortcut around it.

The OECD’s 2026 Digital Education Outlook studied classrooms worldwide and the pattern is clear: the technology isn’t the variable. The teaching is.

So the question isn’t “Will AI replace teachers?”

It’s “Are we using AI to replace teaching, or to support it?”

One boosts performance without learning. The other does both.

Teachers aren’t competing with AI. They’re the reason it actually works.

What’s your take? Have you seen this play out in classrooms?

Source: OECD Digital Education Outlook 2026

No, AI schools aren’t happening. But AI tools in classrooms? Already here.

The future of education isn’t about replacing teachers — it’s about giving them better tools. AI can’t build relationships with students or notice when someone’s struggling. But it can handle the repetitive work that takes hours every week.

Lesson planning. Differentiation. Quiz generation. Material prep.

Teachers are already using AI — not to automate teaching, but to free up time for what actually matters: their students.
We’ve shared practical ways to use AI in the classroom. Like how to plan a week of lessons in 30 minutes, or which tools teachers actually find useful.

Check our recent posts for more. 😉

📷 Images source: Abbott Elementary (ABC)