This important email that you let ChatGPT write, this homework that your teenager completes with a chatbot, this lesson plan generated in one click… In the moment, it’s a relief. Your brain blows, the machine takes care of the rest, and you move on.
But by dint of delegating, a question arises: does AI weaken your critical thinking? Psychologists talk about cognitive unloading, when our brain relies on the tools for a long time. The latest studies describe a very real phenomenon, but also concrete ways to reverse the trend.
AI and cognitive offloading: what really happens to your critical thinking
An article from Psychology Today summarizes the current shift:
“We trade mental effort for convenience, and our brains adapt in alarming ways”writes the magazine. In Switzerland, Michael Gerlich followed 666 people of different ages in 2025. Result: the more frequent the use of generative AI tools, the more critical thinking scores fell on the Halpern Critical Thinking Assessment test. “Cognitive offloading emerged as a mediating factor, particularly among younger participants who exhibited weaker critical thinking skills due to habitual reliance on AI,” explains Michael Gerlich in this study.
Researchers also extend the “Google effect” described in 2011: we mainly remember where to find the information, not the information itself. With chatbots, it’s no longer just the memory that comes out of our head, it’s the entire reasoning. At MIT, an experiment with students showed that just 20% remembered a quote produced with AI, compared to about 85% when they worked without help. Other work, such as that of Wahn and colleagues, shows that “task load is a critical factor in offloading behavior” : the more overwhelmed we feel, the more we let the algorithm think for us.
Seven warning signs that AI is already eating away at your critical thinking
Concretely, cognitive offloading looks like this: you almost always accept the AI’s first response, you have difficulty reformulating a generated text in your own words, you panic when the Wi-Fi cuts out before an assignment or a presentation. Researchers from Microsoft and Carnegie Mellon, working with 319 knowledge workers, observed that the more trust in AI increases, the less people check what it offers, especially for tasks deemed “minor”.
Gerlich also revealed another side: many participants privately say they feel worried about “unlearning” to think, while downplaying this discomfort in public. “Perhaps this anxiety is just a reflection of a recognized truth: AI tools are changing us”notes Psychology Today. Ask yourself three simple questions: Do you rarely try to correct or qualify a chatbot response? Do you quickly forget the generated content? Have you lost the habit of thinking alone for five or ten minutes before opening an AI app? Two positive responses already indicate a shift.
The quick method to strengthen your critical thinking without cutting off the AI
Researchers do not recommend fleeing technology, but taming it. First pillar: install AI-free zones. Fifteen minutes a day where you write a tricky email, analyze an article or solve a problem without any tools. Second pillar: compare. For the same subject, first write down your ideas, then ask the AI for a response and identify what it adds, what it forgets, what seems questionable to you. Third pillar: focus on the “how”. For every important task, keep a record of your reasoning, not just the end result.
Last lever: transform each interaction with AI into a mini-training. Before you click “send”, ask yourself three questions: what could be wrong in this answer, what information is missing, how would you explain it without a screen? Also lighten your mental load when possible, so as not to delegate everything out of fatigue. “The next time you reach out to AI to complete a cognitive task, take a second: which capabilities are you developing, and which are you giving up?” asks Psychology Today.