MIT Study: AI Lowers Brain Activity and Cognition

Published Date : 2/10/2025 

A recent study by MIT has found that the use of AI, particularly ChatGPT, significantly reduces brain activity and cognitive engagement in users, with long-term negative effects on mental performance. 

A study conducted by the Massachusetts Institute of Technology (MIT) has uncovered concerning findings about the impact of AI on human brain activity. The research reveals that when individuals use AI, particularly large language models (LLMs) like ChatGPT, their brains work less hard, and this reduced activity persists, negatively affecting mental performance in future tasks.

The researchers employed a limited number of subjects for their experiments, as detailed in their paper. Participants were asked to write essays on various topics. The subjects were divided into three groups: one group was allowed to use AI (ChatGPT), another was permitted to use Google Search, and the third group, termed ‘brain only,’ produced work without any technological aids.

Electroencephalography (EEG) was used to monitor brain activity and assess cognitive engagement and load. The results showed that the groups exhibited different levels of neural connectivity, reflecting different brain strategies employed during the writing process. The unaided group had the most active grey matter, followed by the search engine group, and the least brain activity was observed among the AI users.

The study also examined ‘ownership,’ the ability of authors to recall and summarize their work. Ownership levels dropped significantly with increased technological assistance. Few students using an LLM could reliably quote what they had written, and the essays produced by the LLM group were more homogeneous, showing less deviation compared to the other groups.

Unsurprisingly, the visual cortex of those using a search engine or ChatGPT was more active, as these groups were more focused on the output of the tools they were using.

To explore longer-term effects, two additional groups were formed: ‘Brain-to-LLM’ and ‘LLM-to-Brain.’ The ‘Brain-to-LLM’ group, which had previously worked without technological aids, now used an LLM. Conversely, the ‘LLM-to-Brain’ group, which had initially used an LLM, was instructed to complete assignments without any technological assistance.

The researchers found that ‘LLM-to-Brain’ participants showed weaker neural connectivity and under-engagement of alpha and beta networks. In contrast, ‘Brain-to-LLM’ participants demonstrated higher memory recall and re-engagement of widespread occipito-parietal and prefrontal nodes. This suggests that using AI after fully exploring one’s thoughts and knowledge can invoke high levels of cognitive integration, memory reactivation, and top-down control.

In summary, humans who use their brains to tackle a subject can benefit from AI after they have fully explored their thoughts, experience, and knowledge. However, those who rely on AI from the outset show reduced brain activity over time and are less capable of performing cognitive tasks without AI assistance.

The study's authors acknowledge the limitations of their research, noting that the sample size was relatively small and that more diverse participants are needed for statistically reliable findings. Nevertheless, the research highlights a pressing concern: the increasing use of AI in educational and everyday settings may lead to a decline in learning skills.

If the trend of using ChatGPT in place of human thinking, considering, and summarizing continues, it is likely that the ability to think effectively will diminish over time. Using AI to add context or additional material later in the intellectual process produces better results than relying on AI from the start.

Search engine use fell between unaided thought and AI-generated content, but the integration of AI-generated results into search engines (such as LLM results appearing at the top of search engine results pages) may lead to a decline in cognitive activity among everyday users if they focus solely on AI-generated content.

The research group emphasizes the need for further study to understand the long-term effects of AI on the brain before LLMs are recognized as beneficial to human cognitive development. 

Frequently Asked Questions (FAQS): 

Q: What was the main finding of the MIT study on AI and brain activity?

A: The main finding of the MIT study was that using AI, particularly ChatGPT, significantly reduces brain activity and cognitive engagement in users, with long-term negative effects on mental performance.

Q: How did the researchers measure brain activity in the study?

A: The researchers used Electroencephalography (EEG) to monitor brain activity and assess cognitive engagement and load in the participants.

Q: What groups were involved in the study?

A: The study involved three groups: one using AI (ChatGPT), another using Google Search, and a third group working without any technological aids.

Q: What are the long-term effects of using AI on cognitive performance?

A: Long-term effects of using AI include reduced brain activity and under-engagement of neural networks, leading to a decline in cognitive performance and the ability to perform tasks without AI assistance.

Q: What are the implications of the study for the use of AI in education?

A: The study suggests that relying heavily on AI in educational settings may lead to a decline in learning skills and cognitive development, emphasizing the need for balanced use of technology in education. 

More Related Topics :