AI can write essays in seconds.
But what happens to the human brain when it does?
A new study from researchers at Massachusetts Institute of Technology explores this question by measuring how people think when writing with and without AI assistance.
Their conclusion raises an interesting possibility:
Using AI too heavily may create something they call “cognitive debt.”
The Experiment
Researchers divided participants into three groups and asked them to write essays.
Each group had different tools:
• Brain-only group – wrote without any assistance
• Search group – could use search engines
• AI group – used an AI writing assistant similar to ChatGPT
While participants wrote, the researchers monitored their brain activity using EEG scans.
The goal was simple: measure how much cognitive effort each group used.
The Brain Activity Gap
The results were striking.
Participants who wrote without tools showed the strongest brain connectivity and cognitive engagement.
Search users showed moderate engagement.
But the group using AI assistants showed the lowest level of brain activity during the writing task.
In other words, when AI helped generate text, the brain appeared to do less work.
Memory and Ownership
The effects went beyond brain activity.
When participants were later asked about their essays, the AI-assisted group struggled more to remember what they had written.
Some even had difficulty quoting sections of their own work.
They also reported a lower sense of ownership over their essays compared with those who wrote independently.
The machine had produced much of the text — and psychologically, the result felt less like their own thinking.
The Idea of “Cognitive Debt”
The researchers describe this phenomenon as cognitive debt.
Just as technical debt accumulates when shortcuts are taken in software development, cognitive debt can build when mental effort is repeatedly outsourced to machines.
AI tools make tasks easier in the moment.
But if overused, they may reduce the mental processes that build skills like:
• reasoning
• memory
• argument development
• critical thinking
The AI Paradox
None of this means AI should not be used.
Tools like ChatGPT are already transforming how people research, write, and work.
The challenge is balance.
AI can accelerate productivity.
But expertise still requires active thinking, not just generated output.
If humans rely on machines to produce ideas, the risk is that the thinking behind those ideas slowly weakens.
For decades, technology has helped people work faster.
Search engines reduced the need to remember facts.
Calculators reduced the need for mental arithmetic.
Generative AI goes one step further.
It can generate entire arguments, essays, and analyses.
That raises a new question:
Not just what AI can do for us.
But what happens to our thinking when it does too much of it.


