MIT Study Finds Using AI Reduces Brain Activity and Ownership

Researchers warn that over-reliance on AI could lead to cognitive decline and apathy.

Feb. 4, 2026 at 7:07pm

Got story updates? Submit your updates here. ›

A recent MIT study found that using large language models like ChatGPT to complete tasks reduces brain activity and cognitive load compared to using one's own brain. Participants who used AI showed less neuronal connectivity and engagement in brain regions associated with mental effort and concentration. The study also found that AI users felt less ownership over their work, raising concerns about the potential for widespread AI use to lead to apathy and a lack of human dignity.

Why it matters

As AI becomes more advanced and integrated into our daily lives, there are growing concerns about the impact on human cognition and critical thinking skills. The study suggests that outsourcing mental tasks to AI could weaken the brain's 'scaffolding' that supports complex reasoning and decision-making. This could have far-reaching implications for education, the workforce, and overall human development.

The details

The MIT study divided participants into three groups - those using only their brains, those using a search tool, and those using ChatGPT - and measured their brain activity and cognitive load while writing essays. The 'brain-only' group showed significantly more neuronal connections and brain activity in regions associated with mental effort, while the ChatGPT group had a much lower cognitive load as the AI handled tasks like developing arguments and structuring the writing. Notably, 16% of the ChatGPT users reported feeling zero ownership over their essays, compared to the brain-only group who felt the strongest sense of ownership.

  • The MIT study was published as a preprint in 2025, but has not yet been peer-reviewed.

The players

Nataliya Kosmyna

The lead scientist on the recent MIT study and a researcher at MIT Media Lab's Fluid Interfaces group and a visiting faculty researcher at Google.

Poppy Crum

A neuroscientist and adjunct professor at Stanford University who was not involved in the recent study.

Lara Boyd

A neuroscientist who has given a TED Talk on how increased difficulty and struggle during practice leads to more learning and greater structural change in the brain.

Got photos? Submit your photos here. ›

What they’re saying

“The bottom line is, you have to do the work. My research has shown that increased difficulty, increased struggle—if you will—during practice actually leads to both more learning and greater structural change in the brain.”

— Lara Boyd, Neuroscientist

“[The homogeneity] was very surprising because some of the participants were not native English speakers. So they put their prompts into ChatGPT, not in English, but in other languages like Spanish. We saw this in the data that the assets would be so similar to each other. And since then, we saw multiple, multiple papers talking about this exact phenomenon about homogenization, very similar language, similar vocabulary used.”

— Nataliya Kosmyna, Lead Scientist, MIT Study

“A lot of people commented on this one specifically saying, well, they haven't written it.”

— Nataliya Kosmyna, Lead Scientist, MIT Study

What’s next

The researchers plan to conduct further studies to better understand the long-term impacts of AI use on human cognition and ownership.

The takeaway

This study highlights the potential risks of over-relying on AI, suggesting it could weaken critical thinking skills and lead to a concerning lack of human agency and dignity. As AI becomes more prevalent, it will be crucial for individuals and society to strike the right balance between leveraging technology and preserving the cognitive benefits of mental effort and struggle.