How to Use AI as a Tutor Without Making Your Brain Lazy

By Emma Kim· October 25, 2026

MIT professor Eric So has a warning about chatbots, and it comes with a fix. Per Fox News, So argues that leaning too hard on AI can weaken memory, problem-solving, and foundational mastery. The reason is that AI removes the mental "friction" your brain needs in order to learn. But he also says the same tools can be valuable when you use them as a tutor or cognitive trainer rather than an answer machine.

This article gives you a simple way to tell the two modes apart, and a few habits that keep you in the helpful one.

Why easy answers can cost you

According to Fox News, So's argument starts with how brains work. AI, he says, takes advantage of the brain's natural tendency to conserve mental effort and seek instant gratification. That creates a feedback loop: the easier it is to ask a chatbot, the more people turn to it, even for the simplest tasks.

The key idea is friction. Learning feels a little uncomfortable when it's working. Trying to recall a fact, wrestling with a stuck problem, or fumbling toward a first draft are the moments when skills get built. If a chatbot jumps in the second things get hard, you get the result without the workout.

That's the hidden cost of instant answers. You still have the finished essay or solved problem, but your own ability hasn't grown.

The two modes: answer machine vs. tutor

So's distinction is the most useful part of the story. Per the report, AI is harmful when used as an answer machine, but valuable as a tutor or cognitive trainer.

Here's a quick way to check which mode you're in:

  • Answer machine: You ask, it produces, you copy. You could not redo the task without it.
  • Tutor: You try first, it responds, you improve. You could explain the result in your own words afterward.

A useful test is to ask yourself, "Could I do a simpler version of this tomorrow without help?" If yes, the AI is probably training you. If no, it may be doing your thinking for you.

Five habits that keep AI in tutor mode

These are practical applications of So's tutor-versus-answer-machine idea. They're our suggestions, not steps from his remarks.

1. Try first, then ask

Before opening a chatbot, spend a few minutes on your own attempt. Write a rough outline, sketch a solution, or jot down what you remember. Even a bad first try keeps the effortful part of learning in your hands.

2. Ask for feedback, not finished work

Instead of "Write this paragraph," try "Here's my paragraph. What's unclear, and what should I fix?" You still do the rewriting, and the AI acts like an editor.

3. Ask it to quiz you

Turn a chatbot into a practice partner. Paste in your notes and ask for questions on them, one at a time, without showing the answers until you respond. Pulling information out of memory is the kind of effort that builds it.

4. Ask for hints before solutions

When you're stuck on a math problem or a piece of code, ask for the next small hint only. If a hint isn't enough, ask for a bigger one. This keeps some of the productive struggle intact while still getting you unstuck.

5. Explain it back

After the AI helps, close the window and explain the idea in your own words, out loud or on paper. If you can't, you've found the gap, and that's useful to know.

When it's fine to hand off the work

Nothing in So's argument, as reported, says you should avoid AI entirely. The problem is outsourcing the skills you're trying to build. For tasks where you don't need to learn anything, such as reformatting a list or tidying a routine email, delegating is a reasonable choice.

The question to ask is whether this is a skill you want to keep or grow. Writing, studying, research, and decision-making are the areas the report highlights, and they're where the habit matters most.

Why this matters for students and workers

AI tools are quickly becoming default helpers for writing, schoolwork, research, and decisions. The conversation is shifting from "Can AI do it?" to "What does AI do to us?" Students and workers face pressure to use these tools constantly, which makes a deliberate approach more useful than a blanket rule.

A practical way to think about it: AI can raise your floor, helping you get unstuck and check your work, but it shouldn't replace your ceiling, the deeper understanding that only comes from doing the hard part yourself.

The takeaway

Next time you reach for a chatbot, make one small change: try the task yourself for five minutes first, then ask the AI to critique your attempt instead of replacing it. That one habit is the difference between a crutch and a coach.