AI Is a Calculator for Your Brain — And You're Losing the Skill

Calculators made us bad at arithmetic. Smartphones made us bad at remembering phone numbers. AI is making us bad at thinking.

This isn’t a prediction — it’s measured. An MIT study that Ali Parandeh references shows the neurological cost of AI dependency. The researchers split people into three groups: one wrote essays using ChatGPT, one used only Google search, and one used their brains and pen-and-paper.

“They measured brain activity,” Parandeh explains. “The group using their brain showed a lot of red lines — very high activity in critical thinking, memory, and communication. The web search group had less. And the ChatGPT group? Maybe one or two red lines. Most of their brain was deactivated.”

This isn’t a moral judgment. It’s a structural problem: as AI takes the cognitive load, your brain’s corresponding circuits go quiet. Use it or lose it.

The Junk Food Parallel

Parandeh’s framework for AI use is oddly practical: “It’s like eating junk food. If you eat junk food once or twice, fine. It’s not going to harm your weight. But if you eat it a lot of times, you start degrading your body. Same thing happens with your brain when you use AI too much.”

The damage isn’t immediate. One essay with ChatGPT doesn’t destroy your critical thinking. One email with Copilot doesn’t cost you your writing instinct. But the pattern matters.

“If you’re firefighting and you use AI to meet a deadline, that’s fine. Or if you’re overloaded with work, use AI to get over that crunch. But you don’t want that to become a repeating pattern. Over time, you lose your intuition.”

Intuition is what you’re actually losing. Not knowledge — intuition. Your subconscious processes millions of data points that feed your judgment. When you stop exercising that judgment, it atrophies.

“You don’t want to lose that intuition because the more you rely on AI, the more you lose it.”

Where It Actually Matters

The calculator metaphor works because calculators came first. We can measure what happened. “Calculators made all of us really bad at arithmetic in our heads,” Parandeh notes. “We have to rely on them now. Sometimes it’s still worth spending time learning a craft manually — doing things that don’t scale in the beginning.”

He learned this by force. When he started making social media content, he had to learn video editing in CapCut. The work didn’t scale. It was slow. But it built something that AI shortcuts skip: craft.

“The writing process helps you crystallize your vision and crystallize your thoughts. If you’re obsessed over quality, that process of writing — even if it takes longer — gives you satisfaction from craft. And you retain the skill.”

This is where the junk food metaphor breaks down. Junk food doesn’t give you satisfaction — it gives you temporary pleasure and lasting damage. But craft does give you satisfaction. The cost of doing it yourself is time. The benefit is competence you keep, intuition you retain, and work that has your signature on it.

The Production Layer

Parandeh runs a solo business. He uses AI extensively: Notion AI, Claude, Figma Make, Copilot 365, CapCut AI. But he’s intentional about where he draws the line.

“If you just get AI to do the entire pipeline for you, you’re not going to get a hundred percent. You’re going to get seventy, eighty percent, maybe if it’s very optimized. Most of the time you won’t get something you like. It won’t have your signature on it.”

He stays in the loop at every stage — Figma to Relume to JetBrains, with him making the call at each handoff. The AI does the heavy lifting. He does the quality control and the decision-making.

“You design the pipeline and get the tools to add the work, but you quality control the outputs between each stage. That’s where you retain the skill and the intuition.”

The pattern he describes is: AI as a layer in a human-controlled system, not a replacement for human judgment. You’re still writing — but AI helps you iterate. You’re still designing — but AI generates options for you to choose from. You’re still building — but AI handles the scaffolding.

The Discipline Part

This requires saying no to convenience. Parandeh’s discipline is specific:

“Any email you send to clients, customers, partners — you always want to double-check with ChatGPT because you become super reliant on these tools. But you don’t want that to happen. You want to cut yourself off. Anytime you feel like you’re becoming too reliant, you need a digital detox from language models. Don’t use Copilot on emails. Just write by hand, even though it wastes more time. You end up not losing your skills. You’re not losing your intuition.”

It sounds tedious. It is. But the alternative is slower degradation — the kind you don’t notice until years in, when you realize you can’t write anything without an AI checking it, or you can’t think through a problem without asking Claude first.

The calculator didn’t go away. We still use them. But we learned to draw the line: calculators for the computation, your brain for the reasoning. AI is the same. The question is where you draw the line — and whether you’re willing to do the slower work yourself in the domains where it matters most.

FAQ

What does the MIT study actually show about AI and brain activity?

Researchers compared three groups writing essays: one used ChatGPT, one used Google search, one used only paper and brain. Brain scans showed the ChatGPT group had minimal brain activity (one or two red lines), the search group had moderate activity, and the brain-only group showed high activation in critical thinking, memory, and communication. The conclusion: AI use correlates with deactivated critical thinking circuits.

Is one essay with ChatGPT going to damage my thinking?

No. One use is fine. The damage comes from the pattern — using AI for every task, every decision, every piece of writing. Like junk food, occasional use doesn’t hurt. Habitual use degraded your skills and intuition over time.

How do I know if I’m overrelying on AI?

Ask yourself: am I reaching for the AI tool immediately, before thinking? Am I unable to write an email without checking it first? Am I avoiding the slower, harder cognitive work? If yes to multiple, you’re in the atrophy zone. Parandeh recommends a digital detox in high-stakes domains like client communication.

What’s the difference between using AI and losing your intuition?

Intuition is your subconscious processing millions of data points into judgment. You build intuition by making decisions repeatedly, seeing feedback, adjusting. When you outsource that to AI, your brain doesn’t build the pattern-recognition muscles. Over time, you stop trusting your own judgment because you’ve atrophied the ability to make it.

Should I avoid using AI for creativity?

Use it, but stay in the loop. Parandeh uses Figma Make for UI generation, Claude for copy iteration, CapCut AI for video editing. But he reviews every output, makes changes, and keeps himself involved. The AI is the tool, not the decision-maker. Your signature should be on the finished work.

How do I use AI without losing critical thinking skills?

Use AI as a layer in a system where you control the flow: generate options, not final answers; iterate, not create once; quality-check outputs between stages. Stay in the feedback loop. Make decisions yourself. Use AI to amplify your thinking, not replace it.

What does “digital detox” from AI actually mean?

For Parandeh, it means deliberately doing cognitive work without AI in high-stakes areas. Write your client email by hand before Copilot touches it. Write your code from scratch before asking for help. Read and think before searching. The detox is the practice. It keeps the skill alive.

Is the calculator comparison actually fair?

Yes. Calculators made us dependent on them for arithmetic, and we lost that skill. We didn’t need that skill as much, so it mattered less. With AI and critical thinking, you still need the skill — especially in judgment-heavy domains like strategy, design, code review, and client communication. The analogy warns you about what’s at risk.

Why does Parandeh’s process — staying in the loop at each stage — matter?

Because staying in the loop means you’re making decisions. You’re evaluating AI output, rejecting what doesn’t work, refining the good parts. That decision-making is where critical thinking happens. It’s the craft part. If you automate the entire pipeline, you miss those moments.

Can I get the MIT study Parandeh mentioned?

He references it as evidence of AI’s effect on brain activation, specifically showing reduced critical thinking circuits in ChatGPT users. The exact study isn’t named in the transcript, but the finding aligns with neuroscience research on skill atrophy from overreliance on external tools.

If I must use AI for productivity, what’s the minimum to stay sharp?

Parandeh suggests: pick one high-stakes domain (writing, design, strategy) and do it without AI once a week. Work slowly. Make mistakes. Think before you ask for help. That one day keeps your skill intact. On the other six, use AI as a force multiplier. That balance prevents atrophy without killing productivity.