Exploring the Impact of AI on Human Creativity and Cognitive Engagement #
Is AI making people less creative? The question has been circling for years, mostly in opinion columns and think pieces. Now the research is catching up. And what it shows is more specific, and more unsettling, than the usual debate about AI replacing jobs.
The concern is not that artificial intelligence will do our work for us. It is that it may be training us to stop thinking altogether.
What the MIT Study Found #
In June 2025, researchers at the MIT Media Lab published the preliminary results of a four-month study examining what happens to the brain when people use AI to write. The study, Your Brain on ChatGPT, used EEG technology to monitor the brain activity of 54 participants divided into three groups: one writing unaided, one using Google Search and one using ChatGPT.
The results were striking. Participants who relied on ChatGPT showed the weakest neural connectivity across all groups, particularly in the frontal theta networks associated with working memory, creativity and executive function. Their writing was more formulaic, less original and ranked lower in quality. They also struggled to recall what they had written just minutes after finishing the task.
By contrast, the brain-only group showed the strongest and most widely distributed neural activity, engaging circuits linked to creativity, language, memory and critical thinking.
When participants switched methods, something equally revealing happened. Those who had relied on AI from the beginning struggled to regain the same level of cognitive engagement when writing unaided. Lead researcher Dr Nataliya Kosmyna described their brains as entering a 'power-saving mode'. The researchers called this gradual erosion 'cognitive debt'.
Beyond Writing #
The MIT study focused on essays, but similar patterns are emerging elsewhere.
A 2025 study published in Societies surveyed 666 participants and found a significant negative correlation between frequent AI use and critical thinking, driven by what the author termed 'cognitive off'. When AI solves the problem, the brain has less work to do. Over time, repeated reliance may weaken independent thinking.
A review published in Frontiers in Psychology found that students who over-relied on AI during creative tasks experienced cognitive fixation and lower creative confidence. Replacing early-stage brainstorming with AI reduced opportunities to explore ideas independently, potentially weakening creative skills over time.
Across the research, the pattern is consistent: AI can increase productivity while reducing the mental effort that develops long-term ability. The work gets done. The person doing it simply engages with it less.
More Than a Classroom Problem #
The implications extend well beyond students and writers.
The same process occurs whenever people outsource decisions that once required thought. Asking AI which sofa to buy instead of developing personal taste. Letting a chatbot solve practical problems instead of working through them. Relying on algorithms to navigate everywhere instead of building a mental map.
Even professional sport has embraced this shift. Football clubs increasingly use algorithms to scout players, reducing the need for extensive in-person observation. Data can identify trends, but it cannot fully capture character, temperament or how a player performs under pressure. Efficiency improves, but human judgement risks becoming secondary.
Individually, these decisions seem insignificant. Collectively, they may reshape how people think, learn and make decisions.
A 2025 randomised controlled experiment published on arXiv found that AI assistance reduced persistence on difficult tasks and left participants performing worse when the technology was removed. Researchers described the process as 'gradual disempowerment'. Users became better at getting answers, but less capable of finding them independently.
The Deeper Problem #
Technology does not have to dominate people to change them. It only has to become habitual.
This is how AI may be influencing human cognition. Not through one dramatic moment, but through countless small delegations of thought. Each decision handed over, each problem avoided and each creative choice deferred to a prompt seems trivial in isolation. Together, they may slowly reshape how people approach thinking itself.
That is why the common reassurance to 'just use AI in the right way' may be less straightforward than it appears. The line between supporting your thinking and replacing it is not fixed. What begins as a useful tool can gradually become the default. The concern is not simply how often people use AI, but whether repeated reliance quietly shifts habits before they even notice.
The Case for Optimism #
The evidence is not entirely one-sided.
Some studies suggest AI can improve creative confidence, generate fresh ideas and enhance collaboration when used to support rather than replace human thinking.
The MIT study itself found that participants who first completed tasks independently before using AI to revise their work achieved the strongest outcomes. They showed stronger brain connectivity, produced better essays and reported higher satisfaction. In other words, AI appeared most useful when it came after the thinking, not before it.
What the Research Actually Says #
The evidence so far suggests AI does not automatically make people less creative. However, habitual reliance on AI to generate ideas, make decisions or complete early-stage thinking appears to reduce the mental engagement that strengthens creativity and critical thinking.
The brain adapts to what it is repeatedly asked to do. Challenge it to generate ideas, evaluate information and solve problems, and those abilities develop. Outsource those tasks consistently, and they may gradually weaken.
The research remains in its early stages. Most studies are relatively small, and the long-term effects across society remain unknown. But the emerging evidence points towards a more nuanced conclusion than either enthusiasts or critics often claim.
AI is neither inherently harmful nor inherently beneficial. Like any tool, its impact depends less on what it can do than on what people stop doing once it is there.
© Copyright IBTimes 2025. All rights reserved.