According to a study conducted at MIT, the use of ChatGPT significantly reduces brain activity.

Your Brain on ChatGPT: Innovation or Cognitive Debt?

Introduction

A groundbreaking study from MIT Media Lab is raising pressing questions about the long-term cognitive effects of AI assisted writing. Titled “Your Brain on ChatGPT: Accumulation of Cognitive Debt”, the research suggests that while tools like ChatGPT enhance speed and efficiency, they may simultaneously erode neural engagement and memory with consequences that linger even after use.

The Science Behind the Findings

• The study involved 54 participants between the ages of 18 and 39, divided into three groups:
• Over four sessions, participants completed writing tasks while their brain activity was recorded using EEG (electroencephalography).
• Neural connectivity, memory recall, and overall writing performance were assessed across the groups.

Key Results

• Reduced Neural Connectivity: ChatGPT users showed significantly lower brain connectivity, particularly in alpha and beta frequency bands indicating diminished neural interaction.
• Memory Decline: 83% of ChatGPT users struggled to recall sentences they had just written, unlike brain-only participants.
• Lasting Effects: Even after stopping ChatGPT use, participants still displayed lower neural interaction compared to non users.
• Quality Concerns: Human evaluators described AI-assisted texts as “robotic,” “shallow,” and lacking depth.
• Productivity vs. Cognition: While ChatGPT users were 60% faster in completing tasks, the mental effort required for genuine learning dropped by 32%.
• Best Performers: Those who began writing unaided and then integrated AI later showed the strongest memory, neural activity, and overall scores.

Limitations and Cautions

• The study is currently a preprint (not peer-reviewed), meaning results are preliminary.
• The participant pool was relatively small (54 individuals) and not fully representative.
• The focus was limited to writing tasks, so findings cannot be generalized across all cognitive domains.
• EEG has limited spatial resolution, restricting insights into deeper brain structures like the hippocampus.

Implications for Education and Work

The findings highlight a paradox: AI can empower users with speed and efficiency, yet it risks diminishing the very cognitive processes that underpin learning and creativity. Over-reliance on AI may accumulate “cognitive debt,” where mental shortcuts today result in reduced mental capacity tomorrow.

However, this does not mean abandoning AI altogether. Rather, it points to the importance of conscious integration:

• Use AI as a support tool, not a replacement for thinking.
• Encourage students and professionals to start with their own ideas, then refine with AI.
• Promote AI literacy so individuals can leverage efficiency without sacrificing depth.

Conclusion

MIT’s study is not the final word, but it sparks a critical conversation about how humans interact with AI. ChatGPT is not inherently harmful the real risk lies in how we use it. When AI replaces mental effort, it may weaken our cognitive engagement. When used deliberately, however, it can amplify human potential.

The future of AI-assisted work will depend not on the tools themselves, but on whether we allow them to guide our thinking or whether we guide them.

Muhammet Furkan BOLAKAR