If you reach for a chatbot for even the simplest tasks, you may want to rethink how you use it. MIT professor Eric So argues that heavy dependence on AI can make human brains less nimble, according to Fox News. The good news: he also says AI can be genuinely valuable for your mind, depending on how you use it. This article turns that distinction into habits you can try today.
What Eric So is warning about
Per Fox News, So says AI exploits the brain's natural tendency to conserve mental effort and seek instant gratification. That creates a feedback loop: the easier it is to get an answer, the more often people turn to chatbots, even for the simplest of tasks.
The concern, as reported, is that overreliance can weaken memory, problem-solving, and foundational mastery. The mechanism is the removal of mental "friction," the effortful struggle that learning seems to need. When a tool smooths that away, you get the answer without building the skill that would let you find it yourself.
The core distinction So draws, per the report, is between two uses:
- AI as an answer machine: harmful, because it skips the thinking.
- AI as a tutor or cognitive trainer: valuable, because it supports the thinking.
That's the whole angle of this piece. The rest is practical: how to land on the tutor side.
Why a little struggle is the point
It's tempting to treat effort as a cost to be minimized. So's argument, as Fox News relays it, flips that. The friction of working a problem is part of how learning happens. Your brain is a bit like a muscle in this respect, though that's our analogy, not a claim from the report.
That's why "productive struggle" matters. If you've wrestled with a math problem, a paragraph that won't come together, or a bug in your code, the solution sticks differently than one handed to you. Skipping that step feels efficient in the moment. The hidden cost shows up later, when you can't do the thing without the tool.
Try this first: attempt before you ask
The simplest tutor-style habit is to put in your own effort before opening the chatbot.
- Take a first pass. Write a rough draft, sketch an outline, or attempt the problem, even badly.
- Then ask for feedback, not output. Instead of "write this for me," try "Where is my reasoning weak?"
- Revise it yourself. Make the changes with your own hands, not by pasting.
This keeps the hard part, the thinking, with you, while still using AI for what a good tutor does: spotting gaps.
Prompts that turn a chatbot into a tutor
These are our suggestions for putting So's tutor-versus-answer-machine idea into practice. They aren't quotes from him.
- "Quiz me on this topic before explaining it." Retrieving what you know exposes what you don't.
- "Give me a hint, not the solution." You get unstuck without losing the struggle.
- "Explain it, then ask me to explain it back." Teaching it back shows whether you really understand.
- "Ask me questions that would help me find the answer." This pushes the reasoning back onto you.
- "Check my work and point out one mistake at a time." Slower, but you do the fixing.
A student-friendly workflow
For schoolwork and studying, try this order of operations:
- Read or work the material on your own first.
- Write down what you think the key ideas are, from memory.
- Ask the AI to quiz you or critique your summary.
- Redo the weak spots without looking at the AI's answer.
- A day later, test yourself again with no tool at all.
That last step is the honesty check. If you can do it unaided, the AI helped you learn. If you can't, it may have just done the work for you.
For work and writing
Professionals face constant pressure to use AI for everything. A few guardrails keep your own skills in shape:
- Decide the thinking yourself. Make the core decision or argument before involving a chatbot, then use it to stress-test.
- Keep some tasks manual. Pick a few routine tasks you do without AI, so the skill doesn't fade.
- Use it for brainstorming after you brainstorm. Generate your own ideas first, then ask for ones you missed.
Why this debate is timely
AI tools are becoming default helpers for writing, schoolwork, research, and decision-making. So the question is shifting from "Can AI do it?" to "What does AI do to us?" That framing is worth carrying into every use. The tool's capability isn't in doubt. What's at stake is whether your own abilities grow or quietly shrink alongside it.
A fair caveat: the Fox News report presents So's argument, and we're relaying it rather than citing new experimental data. Treat the habits above as sensible practice built on his framing, not proven results.
The takeaway
Before you type your next prompt, spend two minutes trying the task yourself. Then ask the AI to coach you, not replace you.
