DEV Community

Tymes AI
Tymes AI

Posted on

AI Is Now Literally Frying Your Brain, And You Are Not Even Realising it

#ai

You can feel it happening in small, everyday moments. You stare at a blank email and feel the pull to let a chatbot write the first sentence. You finish a report with AI's help, and ten minutes later you can barely remember what it said.

People joke that AI is "frying our brains." New research from MIT, Wharton and leading medical journals suggests the joke is closer to a diagnosis than a meme. Relying heavily on AI is weakening memory, dulling instincts and training our minds to switch off.

Convenience was always going to have a cost. We just didn't expect to pay it in thinking.

AI doesn't need to damage brain tissue to leave us mentally sluggish. It works more like a cast on a healthy limb. It takes away the productive struggle that thinking, remembering and problem-solving depend on, and the underlying circuits weaken from lack of use. In classrooms, offices and hospitals, researchers are seeing the same pattern: the more we lean on AI, the less capable we become without it.

AI doesn't make us less intelligent. It makes us less practised. In the end, the result looks much the same.

Your brain powers down when AI takes over

In June 2025, researchers at the MIT Media Lab fitted 54 adults with EEG caps to see what happens in the brain during AI-assisted writing. Over several months, participants wrote 20-minute essays in one of three groups: on their own, with a search engine, or with ChatGPT.
The scans showed a clear decline from one group to the next. People writing unaided showed broad, active communication across brain regions. Search-engine users showed noticeably less. ChatGPT users had the weakest, quietest networks of all.

When the machine does the heavy lifting, the brain stops showing up for work.

The bigger surprise came minutes after the writing stopped. Asked to quote a single line from the essay they had just finished, 83.3% of ChatGPT users (15 of 18) couldn't do it. In the other groups, the figure was 11.1%.

When you prompt an AI and paste its words, your brain skips the effort that turns ideas into memories. Later, when those same ChatGPT users wrote without AI, their brain activity did not return to normal levels, and their writing stayed flat and formulaic.

You can't remember an idea you never had. You can only remember the moment you copied it.

The study is still a preprint and its sample is small, but it matches a pattern neuroscientists have seen before. A long-term McGill University study found that drivers who relied heavily on GPS had poorer spatial memory when navigating without it. Constant digital guidance pulls the brain away from the hippocampus, the region behind memory and mental mapping, and toward autopilot habits. Generative AI is now doing the same to how we reason, write and analyse.

GPS taught us to stop learning the roads. AI is teaching us to stop learning, full stop.

The skill drain: from students to experienced doctors
Defenders of AI point out that it helps people finish faster and produce cleaner work. That's true. But controlled studies suggest the polish comes at a price: our underlying skills are eroding.

Speed is a feature. Dependence is a cost. Most of us are only tracking the first.

At Wharton, researchers gave nearly a thousand high-school maths students access to GPT-4 during practice sessions. Students using an unrestricted chatbot solved far more practice problems. When the AI was removed for the exam, they performed worse than students who had never used it. A second version of the tool, designed to give hints instead of answers, largely avoided the damage.

The students didn't learn maths. They learned to ask for answers.

Medicine shows the same problem, with higher stakes. A 2025 study in The Lancet Gastroenterology & Hepatology followed experienced endoscopists in Poland after AI-assisted colonoscopy was introduced. When these doctors later worked without AI, their rate of detecting precancerous growths fell from about 28% to about 22%. These were not trainees. They were seasoned specialists whose instincts had dulled within months.

If AI can wear down a veteran's expertise in months, nobody's skills should be considered permanent.
This is not an argument against AI.None of this means we should log off and go back to typewriters. AI is one of the most powerful tools ever built. The issue isn't whether we use it. It's what we hand over to it.

The question isn't "Should I use AI?" It's "What am I no longer doing because I use AI?"

The same research points to a better approach. In the MIT study, participants who wrote on their own first and brought in AI afterwards kept stronger brain engagement. At Wharton, students who got hints instead of answers kept their learning. The difference wasn't the technology. It was the order of operations.

Think first, then prompt. The order is the whole strategy.
How to keep your edge

Draft before you prompt. Write your rough take, outline or hypothesis before you open a chatbot. AI should refine your thinking, not replace it.

Ask for challenges, not conclusions. Ask AI to poke holes in your argument, not to make the argument for you.
Keep one AI-free task every day. Write a memo, solve a problem or plan a project entirely in your own head. Treat it like training.

Check what you retained. After any AI-assisted task, try to explain the key points without looking. If you can't, you outsourced the thinking, not just the typing.

Use AI as a tutor, not a ghostwriter. Ask for hints, frameworks and questions. Save full answers for work that doesn't need your judgment.

Use AI as a sparring partner, not a stand-in.

The bottom line

Every major technology has changed how our minds work. Calculators changed arithmetic, search engines changed memory and GPS changed navigation. Generative AI is different because it reaches the faculties we value most: reasoning, writing, judgment and creativity.

The people who thrive won't be the ones who use AI the most. They'll be the ones who stay sharp enough to know when it's wrong.

Top comments (0)