Artificial intelligence was supposed to make us smarter. And in many ways, it has — helping us research faster, analyze complex information and solve problems we couldn’t tackle alone.
How to Use AI Without Losing Your Critical Thinking Skills
- Think before asking AI.
- Use AI as a thought partner.
- Be critical of AI outputs.
- Be cognizant of how you’re using AI.
- Rewrite AI’s outputs in your own words.
But a growing body of research reveals there’s a hidden cost of convenience. Because AI systems can provide a quick, confident answer on nearly every topic imaginable, we may overrely on it to do the thinking for us. And similar to a muscle atrophying from lack of exercise, we will lose our critical thinking skills, our ability to work through hard problems and our motivation to persist when things get difficult.
A new study conducted by researchers at Carnegie Mellon, MIT, Oxford and UCLA posits that signs of this cognitive decay are evident with just 10 minutes of AI usage, and that “sustained AI use risks eroding essential cognitive and motivational capacities, and as a consequence, displacing human participation over time.”
What the Study Found
In a study conducted by researchers at Carnegie Mellon, MIT, Oxford and UCLA, more than 1,200 participants were tested on their reading comprehension skills and fraction-solving abilities. One group was given an AI assistant in the sidebar to help them with the first 10 minutes of the test. Then the AI assistant was taken away, and they were left to solve the last three problems by themselves.
The control group — those without an AI assistant — scored lower during the first portion of the test, but they scored higher once the playing field was leveled. The group who had grown accustomed to AI assistance were also more likely to skip the question, showing that AI use can cause the brain to be less persistent when trying to solve a problem.
The largest declines in performance and persistence were found among the 61 percent of participants who said they used AI to obtain answers. Those declines were not witnessed among participants who said they used AI for hints or clarifications.
“People do not merely become worse at tasks, but they also stop trying,” researchers noted. “If such effects accumulate over months and years of AI use, we may end up creating a generation of learners who have lost the disposition to struggle productively without technological support.”
Growing Evidence of a Problem
This study, which was published in April 2026, is the most thorough effort yet to understand the consequences of users offloading their reasoning and judgment to artificial intelligence. Other studies have found similar results.
In 2025, Michael Gerlich, a professor at the Swiss Business School, surveyed 666 participants and found a strong correlation between AI usage and reduced critical thinking skills. The study found that younger users with higher dependence on AI tools have lower cognitive performance, and that participants with higher levels of education had better critical thinking skills, regardless of their AI usage. Gerlich does not believe that cognitive decay is an inevitable result of AI usage; instead, he sees it as a byproduct of “cognitive offloading” — or delegating mental effort to AI. Users could theoretically use the cognitive resources freed up by AI for innovative tasks, but too often, he told The Big Think, “they channel these resources into passive consumption, driven by AI-enhanced content curation.”
Another study conducted by MIT researchers in 2025 explored how AI usage could lead to “cognitive debt,” or loss of critical thinking skills. The researchers tasked 54 participants with writing essays — one group using ChatGPT, another using Google search and a third group relying solely on their brainpower. Over the course of three writing sessions in a four-month period, the researchers analyzed the brain activity of participants using EEG scans. They found that AI-using participants’ brain connectivity and writing quality was lower than other participants — and that it continued to decrease over time. Moreover, participants were less likely to recall what they had written when they were using ChatGPT.
How AI Can Limit Thinking and Motivation
Concerns about new technologies eroding skills or cognitive abilities are not new. Socrates once warned that writing would impair the memory and active engagement required in the oral tradition. While these concerns are valid, philosopher Walter Ong pointed out that writing has also allowed us to build upon what we’ve learned, leading to more advanced thinking and innovation.
Proponents of AI often compare AI’s current moment to the introduction of calculators in schools in the 1970s and 1980s. Teachers worried that students would no longer learn multiplication tables or long division, which are still taught in schools today. Students still need to understand the basics of arithmetic, but once they understand those concepts, they can use calculators to perform more advanced computations.
Artificial intelligence is not your average calculator, though. It can write essays, make predictions and automate entire processes. Software engineers have fried their brains because the work of agentic coding tools has outpaced their own working memory of the task. Some have stopped analyzing the AI-generated code altogether. So instead of computing numbers like a calculator, AI allows users to arrive at conclusions without understanding the reasoning behind them. If used in that way, students won’t develop the learning tools they need to gain foundational knowledge, let alone build on it.
The fears about technology’s impact on human cognition also surfaced with the rise of Google. Researchers warned that “the Google effect” or “digital amnesia” would lead us to forget information we knew we could find on the internet (or on a digital device) at a later date. And while we may not remember phone numbers or other information like we used to, Google allows us to research any topic imaginable without stepping foot in a library. But unlike search engines that offer a wide range of perspectives, AI assistants typically flatten issues with a single, decisive answer that may be based on bias, false information or a technical misunderstanding. Google users once skimmed several links per search, but now, nearly 65 percent of searches end with users taking AI’s word for it instead of clicking a link.
While AI may be helpful for synthesizing quick information here and there, the curiosity to pursue more information — and the mental persistence to work through a challenge — is essential for skill acquisition and learning. Similar to a frog slowly being burned alive, the researchers at Carnegie Mellon, MIT, Oxford and UCLA warn users may not realize they are losing their capacity for reason, decision-making and judgment until the cumulative effect becomes too overwhelming to overcome.
To address this lack of motivation, researchers argue that, similar to the way a mentor encourages a student to learn by giving hints — withholding information and encouraging increasing levels of autonomy — AI companies should prioritize long-term learning alongside immediate task completion.
5 Ways to Prevent AI From Draining Your Brain Power
While it remains to be seen whether the AI giants will prioritize long-term learning over short-term convenience, here are five practical steps you can take to protect your own mental acuity and motivation in the age of AI.
1. Think Before Asking AI
In MIT’s brain activity study, researchers found that participants who started without AI assistance showed stronger neural connections than people who jumped to AI right away, even after AI was introduced. This suggests that users who genuinely try to solve a problem themselves — even with a rough draft or some preliminary ideas — will preserve more of their mental acuity.
2. Use AI as a Thought Partner
The researchers at Carnegie Mellon, MIT, Oxford and UCLA found the largest declines in performance and persistence were among those who used AI as a means to obtain answers. A more mentally engaging way to use AI is as a thought partner who helps you brainstorm ideas, dig deeper into topics and identify problems with your work.
3. Be Critical of AI Outputs
By now, we should all be aware that AI is prone to bias, sycophancy and other false information. Ask the system to cite evidence for its claims and fact-check any information that feels dodgy. You might also want to ask yourself what the AI could be leaving out and whether there’s a valid counterargument to consider. This critical thinking exercise will keep your cognitive muscles active, and it will likely spare you the embarrassment of regurgitating false information.
4. Be Cognizant of How You’re Using AI
If you’ve noticed that you are using AI to avoid hard work that requires creativity or strategic thinking, sit with the discomfort of not knowing and challenge yourself to work through it — even if it means using AI as a thought partner. At the same time, if you’re wasting time on mundane tasks, don’t be sentimental about automating them to free up time for higher-level work.
5. Rewrite AI’s Output In Your Own Words
The researchers at Carnegie Mellon, MIT, Oxford and UCLA found that essay writers who didn’t have an AI assistant were more likely to retain what they wrote. Even if you relied on AI to help solve a problem, try rewriting the output in your own words without looking at the chat window. This will ensure that you understand the reasoning behind AI’s answer and that your brain was actively engaged in finding a solution.
Frequently Asked Questions
Does AI actually make you less intelligent, or is this just fearmongering?
Research suggests the concern is legitimate. Multiple studies have found that frequent AI use is associated with reduced critical thinking skills, weaker brain activity and lower performance on cognitive tasks — with signs of decline appearing after as little as 10 minutes of use.
Does it matter how I use AI, or is all AI use equally harmful?
How you use AI matters enormously. The largest declines in cognitive performance and persistence were found among people who used AI to get direct answers. People who used AI only for hints or clarifications showed no significant decline.
What is cognitive offloading?
Cognitive offloading is the practice of delegating mental effort to an external tool, such as an AI assistant, instead of doing the thinking yourself. While this can save time and free up mental resources for other tasks, researchers warn that relying on AI for answers too often may weaken critical thinking, problem-solving skills and the motivation to work through difficult challenges on your own.
Who is most at risk of cognitive offloading?
Studies suggest younger users with higher dependence on AI tools show lower cognitive performance. Participants with higher levels of education tended to maintain better critical thinking skills regardless of how much AI they used.
