It was once said that technology frees the human mind to think higher thoughts. But as artificial intelligence (AI) quietly takes over our daily cognitive tasks — from writing emails to finishing our academic assignments — researchers are trying to find the answer to an uncomfortable question: is it also freeing us from thinking altogether?

Recent findings suggest that our increasing reliance on AI tools may be reshaping the way we use, and perhaps lose, our cognitive abilities.

The promise and the paradox

AI’s appeal is irresistible. With a few words, anyone can now generate essays, marketing campaigns, or even code. For students, professionals, and the perpetually busy, AI saves time and effort. But the same convenience that makes AI attractive may come with hidden costs. Researchers call it cognitive offloading — the habit of shifting mental effort to machines.

The Guardian recently analysed a number of studies warning that this offloading could dull critical thinking. Michael Gerlich of the Swiss Business School, who studied over 600 participants in Britain, found that people who relied heavily on AI tools performed worse on critical-thinking assessments. They tended to trust AI-generated information uncritically and struggled to evaluate it independently.

This decline in “mental vigilance” mirrors a wider cultural shift. Tasks that once required sustained thought, for example, reading long texts, structuring arguments, checking sources — are increasingly delegated to algorithms. What we gain in efficiency, we may lose in intellectual engagement.

The MIT study

A recent study by MIT’s Media Lab provides striking neurological evidence of this trend. In an experiment involving 54 students from several Boston universities, researchers tracked brain activity using electroencephalography (EEG) while participants wrote essays in three ways: entirely unaided, with a search engine, or with ChatGPT.

The results were unambiguous. Those who relied on ChatGPT displayed the weakest neural connectivity across the brain’s attention and creative networks. Their brain activity showed less engagement compared with the “brain-only” group, which exhibited the widest and strongest neural patterns.

Even more telling, students who had used ChatGPT in earlier sessions performed worse when later asked to write without it. They struggled to recall details from essays they had just written and expressed a low sense of ownership over their own work.

In contrast, students who began without AI and later introduced it to their writing process showed higher connectivity and better recall. The MIT researchers described this as “accumulated cognitive debt,” which essentially means a decline in mental effort that persists after AI assistance is withdrawn. Over time, dependence on AI could lead to a form of neural efficiency adaptation: the brain, accustomed to outsourcing its labour, learns to do less.

Rethinking intelligence

The implications of these studies extend far beyond the classroom. In workplaces, a study by Microsoft Research found that knowledge workers using generative AI completed tasks with “much less cognitive effort”. The tasks were done faster, but many required little critical review or creative thought.

AI, like calculators or navigation apps, reduces mental workload. But while those tools handle narrow tasks, AI offloads complex reasoning and decision-making. Evan Risko of the University of Waterloo calls this “cognitive miserliness” — the brain’s tendency to choose the easiest path. If left unchecked, AI may reinforce this habit, creating what some researchers call a “feedback loop of mental laziness.”

Yet it would be too simple to say AI makes us stupid. As many experts argue, the technology is only as harmful as the way it is used. When treated as a collaborator rather than a replacement, AI can help refine ideas and stimulate deeper inquiry. Studies show that people who prompt AI step by step, questioning its output rather than accepting it wholesale, retain more control and understanding.

The real challenge, then, is balance. As Dr Barbara Larson of Northeastern University noted, “Long-term critical-thinking decay would likely result in reduced competitiveness.” The smarter approach is to keep AI in its place: a useful but naïve assistant, not a surrogate thinker.

The evidence suggests that every keystroke we surrender to AI slightly reshapes the mind that commands it. Just as calculators once changed how students learned arithmetic, AI is changing how we write, learn, and reason. But unlike calculators, AI touches the very processes that define human creativity and thought.

Perhaps the question is not whether AI is making us stupid, but whether we are allowing it to. The MIT researchers conclude that dependence on AI “had a measurable impact on participants… at neural, linguistic, and behavioural levels.” Their warning is simply this: thinking less may become the easiest habit of all.