At our g2 experience last week, a counselor was telling me about what many are seeing this year. Students feed the same prompt into an AI, and out comes the same essay in slightly different clothes. Her instinct was to stop them from using AI. Mine was different: don’t deny the tool – change how we teach students to think with it.
But the essay is only where it shows up most visibly. The bigger story is happening in classrooms everywhere. Students are asking AI to summarize the chapter before they’ve read it, to solve the math problem before they’ve set it up, to outline the history paper before they’ve decided what matters. The outsourcing isn’t about one assignment – it’s becoming the default habit of school.
AI didn’t invent shortcuts – it just supercharged them. And when shortcuts become the default, students lose practice in the very thing education is supposed to preserve: the ownership of their own thinking.
Psychologists call this cognitive offloading – handing over the hard parts of thinking to someone or something else. We’ve normalized it to the point where students learn to follow directions, not to own decisions.
That’s the danger. When the struggle disappears, so does the practice of judgment. Students stop building the muscles they’ll need when the easy answers run out.
Cognitive offloading didn’t start with AI. We set the stage for it long before.
We built systems that reward correctness over curiosity, speed over struggle, polish over process. We celebrate the neat final product more than the messy journey that produced an honest one. We push students to stay “on track,” then wonder why they freeze when the track ends.
The cost is subtle but profound. When students outsource the hard parts of thinking, they stop practicing the very skills that make thinking meaningful – noticing, judging, revising, making sense of ambiguity. Those muscles of discernment atrophy.
AI just makes the problem impossible to ignore. It’s fluent, fast, and seductive. It can generate an essay about values the student has never lived, or spit out a lab report from data they barely looked at.
But it can also surface perspectives they wouldn’t have found alone, act as a debate partner, push them toward specificity, or reveal contradictions.
We owe students tools, not maps
Our job isn’t to remove all support. It’s to offer the right kind. Not maps that lay out the route, but tools that help them build their own.
Maps tell you where to go. Maps are tidy and reassuring but quickly outdated. Tools are messy, but empowering and durable. When the terrain invariably shifts, the map fails. Tools help you build wherever you are.
AI can be the ultimate map generator – or the sharpest tool in a builder’s hand. Which one it becomes depends on how we design the learning around it.
A better frame for AI
If we insist on purity, we will get performance without ownership. If we insist on ownership, we can still get performance.
Here’s my advice.
- Use AI as a thought partner, not a thought replacement. Ask it for counterarguments, blind spots, analogies, and alternative structures.
- Feed it personal material only the student can supply. Then interrogate what comes back. What feels true, what feels off, what is missing.
- Keep an AI audit trail. Save prompts, drafts, and reflections so the process is visible and gradable. When process counts, students practice it.
In short, we move from answers to agency.
We say we want students to be resilient and creative, yet we rarely put them in situations that actually require either. Builders don’t learn by being told what to do – they learn by doing, then doing again when the conditions change. That’s how we train athletes, coders, musicians, founders. It should also be how we train thinkers.
One way to do this is what I call structured uncertainty. Start with a simple assignment or project. Halfway through, change one meaningful condition: shorten the word count, flip the audience, swap the data set, add a new stakeholder. Now the student has to adapt.
That’s real life. Because in the real world, the plan always changes. The job description shifts. The market turns. The client rewrites the brief. And the people who’ve only ever followed a script? They freeze.
The ones who’ve practiced adapting don’t. They adjust, make a call, and keep moving. That’s the habit we should be building.
The Builder’s Toolkit
Here is a concrete set of practices that counselors and teachers can incorporate into their daily work (no software needed).
- Prompt engineering, for humans
Teach students to ask better questions of AI, mentors, and themselves. Three prompts to master:
- Clarify: “What would a skeptical reader question in my work, and why.”
- Contrast: “Give me three ways to structure this that change the emphasis.”
- Constrain: “Rewrite with one vivid example, two decisions, and no adjectives.”
The point isn’t the prompt. It’s the critical reading of the response. That’s where offloading becomes ownership.
2. The reflection ledger
A short log with every major task: what I tried, what surprised me, what I changed, what I learned. Make it count for a grade. Thinking is part of the work, not extra credit.
3. Three artifacts per assignment
For any big project, collect: raw material, the AI or peer dialogue that shaped the work, and the final product with a short rationale. Shortcutting becomes obvious. Voice becomes visible.
4. Ambiguity trials
Run a 45-minute challenge where the rules change midstream. Score how quickly students adapt, not just the outcome. Debrief what worked.
5. Portfolio thinking
Replace one-and-done with iterative work. Show the trajectory, not just the end result. Progress is a product.
6. Decision memos
At the end of any significant project, have students write one page: options considered, tradeoffs made, what they’d do differently. That’s how output turns into judgment.
7. Ethical use norms
Publish a simple AI policy in student language. What’s encouraged, what must be documented, what crosses the line. Boundaries work better when students helped write them.
From Map-Makers to Tool-Givers
Counselors, teachers, parents – we all fall into the same trap. We smooth the path. We optimize the stops. We hand out checklists and scripts. Students show up, follow the route, and leave with a polished result. Useful? Sure. Transformative? Not really.
- Counselors slip into concierge mode: mapping out course choices, activities, even essays.
- Teachers over-scaffold: spelling out every step, giving templates, grading polish over process.
- Parents over-engineer schedules and decisions, leaving little space for trial and error.
It works in the short term. Students arrive “on time” at the right destination. But when the map no longer matches the ground, they stall.
The alternative is harder, slower, and far more valuable. Let students design their own projects or assignments around problems that matter. Walk alongside them while they figure it out. Help them identify a real user, a real deadline, a real constraint. Push them to collect feedback, adjust, and try again.
That’s when confidence starts to form – not from following instructions, but from making decisions. That’s when small failures become calibration points, and earned successes become identity markers.
And this isn’t just about the strongest kids getting stronger. Students with privilege already get practice navigating gray zones. Students without it are too often handed a rigid script and told to execute it flawlessly. Giving them tools instead of maps is what levels that playing field
The Standards Question
The usual pushback is quality and time. Will this lower standards? Will we have enough hours?
In practice, process grading improves quality because students can no longer fake the middle. They must surface their thinking – where most learning lives. Time is real, so start small. Pilot reflection ledgers. Run one ambiguity trial. Require decision memos on capstones. The first cycle is bumpy. The second is better. By the third, students are teaching us, and each other.
The Stakes
AI will flood the world with competent, pleasant, placeless answers. It will hand out a million free maps to the same destinations. The differentiator will not be who can access the maps. It will be who knows how to build something worth mapping.
If we keep graduating students who only know how to follow directions, we will end up with a society that can’t chart a new course. The students who can ask sharper questions, test assumptions, navigate uncertainty, and choose with integrity will thrive. The rest will wait for instructions, then wonder why the path feels crowded.
We owe students more than directions. We owe them durable tools, honest practice, and the dignity of doing hard things. That is how we preserve cognitive ownership in a time that tempts us to give it away.
If you are a counselor, pick one toolkit item and pilot it with two students this week. If you are a teacher, design one assignment with structured uncertainty and a reflection ledger. If you are a student, keep your audit trail and claim your voice.
The shortcut is optional. Your tools travel with you.
Ex Cogitatione, Progressus.
Girish
