A recent study by the Massachusetts Institute of Technology (MIT) Media Lab on the impact of ChatGPT on the human brain, published in June 2025, has not gone unnoticed by media professionals and media consumers – and that is rightly so. MIT neuroscientists conducted a study on 54 volunteer MIT undergraduates and postdoctoral researchers (aged 18 to 39), who were required to write essays with and without the assistance of ChatGPT to assess its influence on their neural activity (Belot, 2025). In order to meet this goal, they used EEG headsets that effectively measure and interpret brain processes. They found that using ChatGPT significantly weakens brain engagement, reducing activity in the networks of focus, memory and attention (Kosmyna, 2025). Such results are alarming for society at a broader level, especially considering the evolution and improvement of AI on a yearly basis, and people’s dependence on it to perform even simple everyday tasks.
The MIT participants were divided into three different groups: The first group was instructed to rely solely on the Large Language Model (LLM) – ChatGPT; the second used Search Engines like Google; while the third group worked without any tools, relying only on their own knowledge. The researchers monitored neural activity with a non-invasive brain imaging technique known as EEG. Results indicate that a noticeable difference emerges when utilising tools like LLMs compared to writing with search engines or without any aids. Among the three groups, those using ChatGPT exhibited the least cognitive involvement. ChatGPT users tended to become increasingly indolent with each subsequent essay, frequently opting for copy-and-paste methods (Kosmyna, 2025).
Following the third session, researchers called the participants back for a fourth time. Only 18 out of the original 54 agreed to return, resulting in a reduced sample size (Belot, 2025). This time around, the students were organised into new groups – the brain-only writers were now assigned to use ChatGPT. This change revealed an unexpected finding. Participants who had engaged in brain-only sessions for the initial three essays achieved even better results with ChatGPT. Repeated use of their cognitive skills on the assigned tasks, when given an additional tool in the next session, made them more inclined to be critical and ask the right questions (Kosymna, 2025; Chow, 2025).
Nataliya Kosmyna, research scientist at the MIT Media Lab and lead author of the paper, “felt it was important to release the findings to elevate concerns that as society increasingly relies upon LLMs for immediate convenience, long-term brain development may be sacrificed in the process” (Chow, 2025). Biologist Ron Brooks, speaking on the podcast “How Artificial Intelligence Could Affect Human Evolution,” compared the relationship between humans and artificial intelligence tools to the mutualism that occurs in nature, where one subject benefits from another and vice versa. Nonetheless, this type of mutualism can “morph into parasitism”, ending in harmful interactions and consequences between the parties. And this is exactly what is happening today, as research shows.
Producing a diverse range of content, including text, images, videos, audio, and computer codes, takes away our chance to practice these tasks and sharpen our cognitive skills, diminishing the neural infrastructure that underpins them. Similar to how neglecting physical exercise can lead to muscle loss, relying on cognitive shortcuts can weaken neural connections. This phenomenon was highlighted in a study in the UK involving 666 AI users, which revealed a notable link between regular Artificial Intelligence usage and diminished critical thinking abilities (Gerlich, 2025) . Additionally, a survey of 319 professionals in Pennsylvania who engage with GenAI weekly showed that while it enhanced their productivity, it also impaired their critical thinking skills and encouraged a long-term dependence on technology. Researchers warn that this could lead to a decreased capacity to tackle problems independently of AI assistance (Lee at al, 2025, p.1-3).
People’s lives have encountered worrying changes throughout history as a result of emerging technologies; for instance, mobile phones and social media networks provide various advantages like communication and access to information, yet they are linked to adverse effects on both physical and mental wellbeing, along with social and cognitive abilities. These consequences stem from how they are utilised and incorporated into our routines. Similarly, AI, which was originally created and developed to help users efficiently find answers and services, is now being criticised for undermining human intelligence and creativity.
Holding AI and its technologies entirely responsible would be unfair. It is the irresponsible application and usage, the lack of AI education, policies and legislations that enable AI to pose a risk to human development. Of high importance is also testing these tools before implementing them. Governments should enable ethical AI while reducing the extensively discussed risks and helping society thrive. Digital global citizens and institutions should come to terms with AI being a permanent fixture and figure out how to engage with it appropriately; otherwise, we risk not only becoming obsolete but also diminishing our higher-order thinking capacities. Accordingly, the interaction with it demands the same competencies that a media-literate individual should have and apply when interacting with any media service. First and foremost, it is essential to actively question, judge and evaluate the information generated or the tasks automated for them. Users of AI technologies should be conscious of the fact that by utilising and blindly consuming AI content, we give those in control of it more power over us and our future.
References
Kosmyna, N. (2025). Project Overview ‹ Your Brain on ChatGPT – MIT Media Lab.
Ware, G. (2025). How AI could influence the evolution of humanity – podcast. The Conversation.
Related Posts
A recent study by the Massachusetts Institute of Technology (MIT) Media Lab on the impact of ChatGPT on the human brain, published in June 2025, has not gone unnoticed by media professionals and media consumers – and that is rightly so. MIT neuroscientists conducted a study on 54 volunteer MIT undergraduates and postdoctoral researchers (aged 18 to 39), who were required to write essays with and without the assistance of ChatGPT to assess its influence on their neural activity (Belot, 2025). In order to meet this goal, they used EEG headsets that effectively measure and interpret brain processes. They found that using ChatGPT significantly weakens brain engagement, reducing activity in the networks of focus, memory and attention (Kosmyna, 2025). Such results are alarming for society at a broader level, especially considering the evolution and improvement of AI on a yearly basis, and people’s dependence on it to perform even simple everyday tasks.
The MIT participants were divided into three different groups: The first group was instructed to rely solely on the Large Language Model (LLM) – ChatGPT; the second used Search Engines like Google; while the third group worked without any tools, relying only on their own knowledge. The researchers monitored neural activity with a non-invasive brain imaging technique known as EEG. Results indicate that a noticeable difference emerges when utilising tools like LLMs compared to writing with search engines or without any aids. Among the three groups, those using ChatGPT exhibited the least cognitive involvement. ChatGPT users tended to become increasingly indolent with each subsequent essay, frequently opting for copy-and-paste methods (Kosmyna, 2025).
Following the third session, researchers called the participants back for a fourth time. Only 18 out of the original 54 agreed to return, resulting in a reduced sample size (Belot, 2025). This time around, the students were organised into new groups – the brain-only writers were now assigned to use ChatGPT. This change revealed an unexpected finding. Participants who had engaged in brain-only sessions for the initial three essays achieved even better results with ChatGPT. Repeated use of their cognitive skills on the assigned tasks, when given an additional tool in the next session, made them more inclined to be critical and ask the right questions (Kosymna, 2025; Chow, 2025).
Nataliya Kosmyna, research scientist at the MIT Media Lab and lead author of the paper, “felt it was important to release the findings to elevate concerns that as society increasingly relies upon LLMs for immediate convenience, long-term brain development may be sacrificed in the process” (Chow, 2025). Biologist Ron Brooks, speaking on the podcast “How Artificial Intelligence Could Affect Human Evolution,” compared the relationship between humans and artificial intelligence tools to the mutualism that occurs in nature, where one subject benefits from another and vice versa. Nonetheless, this type of mutualism can “morph into parasitism”, ending in harmful interactions and consequences between the parties. And this is exactly what is happening today, as research shows.
Producing a diverse range of content, including text, images, videos, audio, and computer codes, takes away our chance to practice these tasks and sharpen our cognitive skills, diminishing the neural infrastructure that underpins them. Similar to how neglecting physical exercise can lead to muscle loss, relying on cognitive shortcuts can weaken neural connections. This phenomenon was highlighted in a study in the UK involving 666 AI users, which revealed a notable link between regular Artificial Intelligence usage and diminished critical thinking abilities (Gerlich, 2025) . Additionally, a survey of 319 professionals in Pennsylvania who engage with GenAI weekly showed that while it enhanced their productivity, it also impaired their critical thinking skills and encouraged a long-term dependence on technology. Researchers warn that this could lead to a decreased capacity to tackle problems independently of AI assistance (Lee at al, 2025, p.1-3).
People’s lives have encountered worrying changes throughout history as a result of emerging technologies; for instance, mobile phones and social media networks provide various advantages like communication and access to information, yet they are linked to adverse effects on both physical and mental wellbeing, along with social and cognitive abilities. These consequences stem from how they are utilised and incorporated into our routines. Similarly, AI, which was originally created and developed to help users efficiently find answers and services, is now being criticised for undermining human intelligence and creativity.
Holding AI and its technologies entirely responsible would be unfair. It is the irresponsible application and usage, the lack of AI education, policies and legislations that enable AI to pose a risk to human development. Of high importance is also testing these tools before implementing them. Governments should enable ethical AI while reducing the extensively discussed risks and helping society thrive. Digital global citizens and institutions should come to terms with AI being a permanent fixture and figure out how to engage with it appropriately; otherwise, we risk not only becoming obsolete but also diminishing our higher-order thinking capacities. Accordingly, the interaction with it demands the same competencies that a media-literate individual should have and apply when interacting with any media service. First and foremost, it is essential to actively question, judge and evaluate the information generated or the tasks automated for them. Users of AI technologies should be conscious of the fact that by utilising and blindly consuming AI content, we give those in control of it more power over us and our future.
References
Kosmyna, N. (2025). Project Overview ‹ Your Brain on ChatGPT – MIT Media Lab.
Ware, G. (2025). How AI could influence the evolution of humanity – podcast. The Conversation.
Related Posts
A recent study by the Massachusetts Institute of Technology (MIT) Media Lab on the impact of ChatGPT on the human brain, published in June 2025, has not gone unnoticed by media professionals and media consumers – and that is rightly so. MIT neuroscientists conducted a study on 54 volunteer MIT undergraduates and postdoctoral researchers (aged 18 to 39), who were required to write essays with and without the assistance of ChatGPT to assess its influence on their neural activity (Belot, 2025). In order to meet this goal, they used EEG headsets that effectively measure and interpret brain processes. They found that using ChatGPT significantly weakens brain engagement, reducing activity in the networks of focus, memory and attention (Kosmyna, 2025). Such results are alarming for society at a broader level, especially considering the evolution and improvement of AI on a yearly basis, and people’s dependence on it to perform even simple everyday tasks.
The MIT participants were divided into three different groups: The first group was instructed to rely solely on the Large Language Model (LLM) – ChatGPT; the second used Search Engines like Google; while the third group worked without any tools, relying only on their own knowledge. The researchers monitored neural activity with a non-invasive brain imaging technique known as EEG. Results indicate that a noticeable difference emerges when utilising tools like LLMs compared to writing with search engines or without any aids. Among the three groups, those using ChatGPT exhibited the least cognitive involvement. ChatGPT users tended to become increasingly indolent with each subsequent essay, frequently opting for copy-and-paste methods (Kosmyna, 2025).
Following the third session, researchers called the participants back for a fourth time. Only 18 out of the original 54 agreed to return, resulting in a reduced sample size (Belot, 2025). This time around, the students were organised into new groups – the brain-only writers were now assigned to use ChatGPT. This change revealed an unexpected finding. Participants who had engaged in brain-only sessions for the initial three essays achieved even better results with ChatGPT. Repeated use of their cognitive skills on the assigned tasks, when given an additional tool in the next session, made them more inclined to be critical and ask the right questions (Kosymna, 2025; Chow, 2025).
Nataliya Kosmyna, research scientist at the MIT Media Lab and lead author of the paper, “felt it was important to release the findings to elevate concerns that as society increasingly relies upon LLMs for immediate convenience, long-term brain development may be sacrificed in the process” (Chow, 2025). Biologist Ron Brooks, speaking on the podcast “How Artificial Intelligence Could Affect Human Evolution,” compared the relationship between humans and artificial intelligence tools to the mutualism that occurs in nature, where one subject benefits from another and vice versa. Nonetheless, this type of mutualism can “morph into parasitism”, ending in harmful interactions and consequences between the parties. And this is exactly what is happening today, as research shows.
Producing a diverse range of content, including text, images, videos, audio, and computer codes, takes away our chance to practice these tasks and sharpen our cognitive skills, diminishing the neural infrastructure that underpins them. Similar to how neglecting physical exercise can lead to muscle loss, relying on cognitive shortcuts can weaken neural connections. This phenomenon was highlighted in a study in the UK involving 666 AI users, which revealed a notable link between regular Artificial Intelligence usage and diminished critical thinking abilities (Gerlich, 2025) . Additionally, a survey of 319 professionals in Pennsylvania who engage with GenAI weekly showed that while it enhanced their productivity, it also impaired their critical thinking skills and encouraged a long-term dependence on technology. Researchers warn that this could lead to a decreased capacity to tackle problems independently of AI assistance (Lee at al, 2025, p.1-3).
People’s lives have encountered worrying changes throughout history as a result of emerging technologies; for instance, mobile phones and social media networks provide various advantages like communication and access to information, yet they are linked to adverse effects on both physical and mental wellbeing, along with social and cognitive abilities. These consequences stem from how they are utilised and incorporated into our routines. Similarly, AI, which was originally created and developed to help users efficiently find answers and services, is now being criticised for undermining human intelligence and creativity.
Holding AI and its technologies entirely responsible would be unfair. It is the irresponsible application and usage, the lack of AI education, policies and legislations that enable AI to pose a risk to human development. Of high importance is also testing these tools before implementing them. Governments should enable ethical AI while reducing the extensively discussed risks and helping society thrive. Digital global citizens and institutions should come to terms with AI being a permanent fixture and figure out how to engage with it appropriately; otherwise, we risk not only becoming obsolete but also diminishing our higher-order thinking capacities. Accordingly, the interaction with it demands the same competencies that a media-literate individual should have and apply when interacting with any media service. First and foremost, it is essential to actively question, judge and evaluate the information generated or the tasks automated for them. Users of AI technologies should be conscious of the fact that by utilising and blindly consuming AI content, we give those in control of it more power over us and our future.
References
Kosmyna, N. (2025). Project Overview ‹ Your Brain on ChatGPT – MIT Media Lab.
Ware, G. (2025). How AI could influence the evolution of humanity – podcast. The Conversation.
