In a study by MIT (Massachusetts Institute of Technology), the human brain is not only less effective when using LLM, but its effects continue to affect mental activity in future work.
The researchers used a limited number of subjects in the experiment (limitations listed in the paper (PDF)). One group of subjects was allowed to use AI (CHATGPT was chosen). Researchers were allowed to generate work without the second group allowed to use Google search, i.e. “Brain Only”, i.e., without technology AIDS.
Electroencephalography (EEG) was used in all subjects to monitor brain activity to promote cognitive involvement and load. Researchers found that the group exhibited different levels of neural connectivity. This reflects the various strategies employed by the brain to write up assignments. The more support the subjects provided, the less their brains appeared to work. EEG analysis showed that the most active gray matter belongs to the auxiliary group, with less neural activity in the “search engine group” and less neural activity, at least among AI users.
The study also looked at what is called “ownership.” This is the ability to summarise the work by citing what the author wrote later. The level of ownership declined dramatically, and subjects received from the technology were more helpful. Few students using LLM were able to reliably quote what they wrote. Furthermore, the LLM-used group “generates statistically homogeneous essays within each topic, indicating significantly less deviations compared to the other groups.”
Naturally, the visual cortex of those using search engines or ChatGPT is more active, and these groups say they want to “focus on the output of the tools they were using.”
Long-term impact
After several essay writing, two more groups were formed from participating subjects, consisting of “Lm from the brain” and “Lm from the brain.” This was a subject who had no technology AIDS that previously failed to use LLM, and LLM users previously instructed them to complete the task.
The researchers stated, “Participants from LLM to the brain showed weaker neural connectivity and underengagement of the alpha and beta networks. Participants from the brain to LLM showed higher memory recalls and extensive occipital and prefrontal remarriages. (…) top-down control.”
In short, humans can benefit from using AI using their brains to tackle subjects rear They already have a complete grasp of their thoughts, experiences, knowledge and emotions without using technology. However, those using AI from the start were unable to perform cognitive tasks when brain activity decreased over time and were asked to use ChatGpt.
The paper states, “As demonstrated over a four-month period, participants in the LLM group performed worse than their counterparts in the brain-only group at all levels: neural, language, and (and) scoring.”
Limited research
The study had only a few dozen subjects, so the research group used a limited sample. The authors acknowledge that it is necessary to use more volunteers from more diverse backgrounds to reveal more statistically reliable findings. However, as AI is increasingly used in schools, universities and in everyday life, researchers highlight what is called the “urgent problem” of “decreased learning skills” that results from using AI as an alternative to the human brain.
Conclusion
If the tendency to use ChatGpt instead of highly human thinking activities continues, it appears likely that the ability to think effectively will decrease over the long term. When AI adds context or additional material, it gives better results later in any process of intellectual consideration than it would be used from the start.
According to the paper, the use of search engines fell into the middle ground between independent thinking and being a spoon-grown AI-generating material. However, Google and Microsoft needs, et al. Inserting AI-Generation into a user’s search results (LLM results (search engine results page) that appears at the top of the SERP) means that cognitive activity among everyday search users can be reduced if AI focuses solely on the search results.
The research group states that understanding the long-term effects of AI on the brain requires more research before LLM is recognized as a net positive for humans.
(Image Source: “Cognitive Test” by Nestle is licensed under CC BY-NC-ND 2.0.))
Want to learn more about AI and big data from industry leaders? Check out the AI & Big Data Expo in Amsterdam, California and London. The comprehensive event is part of TechEx and will be held in collaboration with other major technology events. Click here for more information.
AI News is equipped with TechForge Media. Check out upcoming Enterprise Technology events and webinars here.
