A reflection inspired by the MIT study “Your Brain on ChatGPT.”

Recently, I started reading a large number of academic papers for my courses. While doing this, I became very curious about how we read and how we write. Should I search using Google and read original papers? Or should I rely on AI summaries?

Then I came across an MIT study called “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task.”
The researchers divided participants into three groups:

ChatGPT group
Google Search group
Brain-only group (no tools, only thinking)

What the study found shocked me

The group that used ChatGPT mostly copied and pasted answers. Their brain activity was the lowest, and many didn’t really understand what they wrote.
The EEG data showed that their brains were less engaged and less connected.

The brain-only group showed the highest cognitive engagement, deeper understanding, and stronger neural connections. Afterwards, they also reported higher satisfaction and a stronger sense of achievement.

The Google Search group was in between:
still active thinking, but supported by external information.

The surprising part: role reversal

Later, MIT researchers swapped the groups.

The original ChatGPT group could no longer use AI.
But these participants discovered something scary:

They couldn’t remember what they had written earlier.
Because ChatGPT had done the thinking for them, their brains never encoded the knowledge.

Their memory formation decreased. Their cognitive processing weakened.
The paper literally says they accumulated “cognitive debt.”

This made me realize something:
Are we turning ourselves into workers on an AI assembly line?
Copy → paste → submit → repeat.
Fast, but empty.

Efficiency is not true learning

Yes, AI is fast.
But if you rely on it blindly, you never build your own neural pathways.
You never strengthen your memory.
You never internalize knowledge.

Eventually, you become mentally weaker, not smarter.

Why I Read Original Papers

My current method is:

read in the original language (English)
look up terms using a dictionary or Google
avoid AI summaries unless necessary

Not because I have extra time,

but because reading the original text helps me understand the author’s intention, tone, and reasoning.
It also strengthens my vocabulary in my research field, helping me:

speak confidently at international conferences
use domain-specific terminology
think critically
identify gaps, challenges, and breakthroughs in the field
communicate smoothly with native English speakers

AI may save time, but wrong usage destroys thinking.

The conclusion is simple:
AI isn’t the enemy — lazy usage is.
In the age of AI, we must be even more disciplined and intentional.

My Question

In the AI era, can we still preserve our critical thinking?

From what I see around me, students usually fall into two groups:

Rational Users (GPT + Research)

They use AI for support, but still read, analyze, and think deeply.

Copy-Paste Users (Only GPT)

They don’t understand what they write.
They skip original texts.
They want shortcuts.
And when asked to think critically, they all give the same generic answers.

Why?

Because their brains never activated the neural connections required to analyze and question ideas.
Their minds have no “raw material” to think with.

The real challenge for us students

We must learn to balance:

Time efficiency
and
Quality of thinking

This balance determines whether we remain independent thinkers
or become passive operators waiting for AI output.

The MIT study is a strong reminder:

If you don’t think, your brain stops learning.
If you stop learning, AI will think for you — and eventually instead of you.

AI can help us, but it should never replace our minds.

MIT study: https://arxiv.org/pdf/2506.08872