Photo by Mikhail Nilov on Pexels

New research from MIT suggests that relying on Large Language Models (LLMs) like ChatGPT could have lasting effects on cognitive function. The study, which used EEG to monitor brain activity, found that individuals who used AI assistance to write essays exhibited reduced neural activity compared to those who used Google Search or worked independently.

The research participants were divided into three groups: one using ChatGPT, another using Google Search, and a control group working independently. Those using AI showed significantly less brain activity. Furthermore, they demonstrated a diminished ability to recall and summarize their own writing, even after the AI tool was removed.

Interestingly, when participants who initially used AI were later asked to perform tasks without assistance, they exhibited weaker neural connectivity than those who had initially worked independently. Conversely, those who initially worked independently saw cognitive benefits from using AI in subsequent tasks. These findings suggest that while AI can enhance cognitive abilities after initial independent thought, early reliance may hinder cognitive development. The study also found that search engine use occupied a middle ground in terms of impact on brain activity. Researchers emphasize the need for further investigation into the long-term cognitive consequences of AI integration, particularly as AI becomes increasingly embedded within search engines.