Researchers at the Massachusetts Institute of Technology have raised concerns about the growing use of generative AI in daily cognitive tasks, following a study that examined how people interact with large language models when writing essays. According to findings from the MIT Media Lab, participants who leaned on tools like ChatGPT for support demonstrated a noticeable drop in brain activity during the writing process, particularly when compared to those relying solely on their own thinking or traditional internet searches.
The study, which involved just over 50 participants between the ages of 18 and 39, divided the group into three categories to test different methods of completing a writing assignment. One group completed the task without digital help, another used conventional search engines, and the third was allowed to work with AI systems built on large language models. While those who used AI finished the task with less frustration and effort, they also showed less mental engagement, based on measurements recorded during the exercise.
Although the results are not entirely unexpected, they add to a growing body of research exploring how AI tools may influence human behaviour over time. Participants who depended on AI support were less likely to apply analytical thought when producing their responses. When asked to repeat the same essay task without any help, their performance tended to drop, which suggested a kind of cognitive decline linked to earlier reliance on machine-generated assistance.
Beyond the drop in brain activity, the researchers noted that participants working without AI support were more thoughtful about ethical implications related to the task, indicating deeper personal engagement with the material. Meanwhile, those using AI showed a tendency to accept generated responses at face value, without questioning the accuracy or intent behind them.
While the study itself was limited in scope and has yet to undergo formal peer review, it joins a larger discussion about how overreliance on automated systems might gradually reduce the need for independent thinking. Similar concerns were echoed in earlier research conducted by Microsoft and Carnegie Mellon University, which likened the decline in mental effort to the weakening of unused muscles. There is also increasing scrutiny from educators and policymakers, particularly given that students have been documented turning to AI tools as shortcuts for academic work.
Although digital assistants may offer convenience and speed, especially for repetitive or time-consuming tasks, the long-term trade-offs remain uncertain. The MIT researchers pointed out that more extensive studies will be needed to understand how repeated cognitive offloading to AI systems might shape learning, reasoning, and brain function in the future. Until more evidence is available, the findings act as an early warning that depending too much on AI may carry more risk than initially assumed.
Image: DIW-Aigen
Read next: Inside the Chat: Could WhatsApp Be Hacked by a Government? Expert Reveals How
The study, which involved just over 50 participants between the ages of 18 and 39, divided the group into three categories to test different methods of completing a writing assignment. One group completed the task without digital help, another used conventional search engines, and the third was allowed to work with AI systems built on large language models. While those who used AI finished the task with less frustration and effort, they also showed less mental engagement, based on measurements recorded during the exercise.
Although the results are not entirely unexpected, they add to a growing body of research exploring how AI tools may influence human behaviour over time. Participants who depended on AI support were less likely to apply analytical thought when producing their responses. When asked to repeat the same essay task without any help, their performance tended to drop, which suggested a kind of cognitive decline linked to earlier reliance on machine-generated assistance.
Beyond the drop in brain activity, the researchers noted that participants working without AI support were more thoughtful about ethical implications related to the task, indicating deeper personal engagement with the material. Meanwhile, those using AI showed a tendency to accept generated responses at face value, without questioning the accuracy or intent behind them.
While the study itself was limited in scope and has yet to undergo formal peer review, it joins a larger discussion about how overreliance on automated systems might gradually reduce the need for independent thinking. Similar concerns were echoed in earlier research conducted by Microsoft and Carnegie Mellon University, which likened the decline in mental effort to the weakening of unused muscles. There is also increasing scrutiny from educators and policymakers, particularly given that students have been documented turning to AI tools as shortcuts for academic work.
Although digital assistants may offer convenience and speed, especially for repetitive or time-consuming tasks, the long-term trade-offs remain uncertain. The MIT researchers pointed out that more extensive studies will be needed to understand how repeated cognitive offloading to AI systems might shape learning, reasoning, and brain function in the future. Until more evidence is available, the findings act as an early warning that depending too much on AI may carry more risk than initially assumed.
Image: DIW-Aigen
Read next: Inside the Chat: Could WhatsApp Be Hacked by a Government? Expert Reveals How
