Is AI Making You Dumber? What the Research Actually Says

You asked AI to draft the email. Then to summarize the report. Then to fix the code, plan the trip, and settle an argument you were losing. It was fast, and honestly it felt great.

Then a quieter question shows up: if it does all the thinking, what happens to mine?

It's a fair worry, and researchers have started putting numbers on it. Here's what they've found so far, what they haven't, and how to keep using AI without going soft.

The short answer

Yes, relying too heavily on AI can weaken specific thinking skills. But the tool isn't the problem. How you use it is.

What the studies show

Heavy AI users tend to score lower on critical thinking. In a 2025 study of 666 people, Michael Gerlich of SBS Swiss Business School found that more frequent AI use went hand in hand with weaker critical-thinking scores. The study appeared in the journal Societies. The driver seemed to be a habit psychologists call cognitive offloading: letting an outside tool do mental work you'd otherwise do yourself. Younger users relied on AI the most and scored the lowest. People with more education scored better on critical thinking no matter how much AI they used.

Trusting AI too much can switch off your own judgment. Microsoft and Carnegie Mellon University surveyed 319 knowledge workers about using generative AI at work, and presented the results at the CHI 2025 conference. The finding that stuck with me: people who trusted the AI more reported doing less critical thinking, while people who trusted their own skills more reported doing more of it. Put plainly, the more you lean on the machine, the less you bother to check it.

Specific skills can fade, but your core brain probably doesn't. A theme runs through this research. When you hand a task to AI, the particular skill behind that task gets rusty because you stop practicing it. That applies to writing, coding, or building an argument. Foundational abilities like attention and working memory seem to hold up better. You don't forget how to think. You just lose your touch at the things you've quietly stopped doing.

The honest caveats

Before anyone panics, this is where a lot of headlines overreach, so a few things are worth knowing.

• A link is not proof. The 666-person study shows a correlation, not cause and effect. Maybe AI dulls thinking. Or maybe people who already avoid hard thinking are the ones who reach for AI most. The study can't tell us which.
• A lot of the data is self-reported. The Microsoft research measured how much critical thinking people felt they did, through a survey. That isn't the same as testing their performance. Other studies lean on self-reports or narrow groups too, so sweeping conclusions are premature.
• Motivation matters more than volume. The wider literature keeps landing on the same point: why you use AI matters more than how often. People who use it to dig deeper come out ahead of people who use it to dodge the work.

The evidence is real. It's also early, and it does not say AI makes everyone dumber.

Why it happens

None of this is new. We stopped memorizing phone numbers once our contacts lived on our phones. Plenty of us can't find our way across a new city without GPS. Whenever a tool takes over a task, we skip the practice that kept us good at it.

What's different with AI is what it takes over. A calculator handles arithmetic, which is a narrow slice of thinking. AI reaches into reasoning, writing, and judgment, which is most of thinking. That's the part that makes this feel higher-stakes than a calculator ever did.

How to use AI without losing your edge

1. Think first, prompt second. Take a real shot at the problem, or write a rough draft, before you ask AI anything. Then let it critique or sharpen what you made. You keep the reasoning and it polishes the result.
2. Ask it to teach, not just tell. Instead of a finished answer, ask for the explanation, the counterargument, or a quick quiz to see if it actually stuck.
3. Verify before you trust. Treat the output like a draft from a clever colleague who is sometimes wrong. Check the facts and the logic, because these tools state nonsense with total confidence. Our ChatGPT vs Claude vs Gemini comparison digs into how each one handles accuracy.
4. Guard the skills you care about. Want to stay a strong writer or coder? Keep doing some of that work by hand. Skills are like muscles, and this is the gym.
5. Stay curious. Motivation is the thread the research keeps returning to. Use AI to learn more, not to learn less.

The bottom line

AI is a genuinely powerful tool. Used well, it can make you both faster and sharper. Used as a stand-in for thinking, it quietly wears down the skills you lean on most.

The point was never to avoid AI. It's to stay the one doing the thinking. If you're folding it into your day-to-day, our guide to the best AI tools that actually save time is about handing it the busywork and keeping the judgment for yourself.

Sources

• Michael Gerlich, Societies (MDPI), "AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking" — 2025 link
• Lee et al., Microsoft Research and Carnegie Mellon (CHI 2025), "The Impact of Generative AI on Critical Thinking" — 2025 link
• Microsoft Research Blog, "The Future of AI in Knowledge Work: Tools for Thought at CHI 2025" — 2025 link

What's your rule for using AI without over-relying on it?