techend
By Ariz Riaz :
A recent study from the Massachusetts Institute of Technology (MIT) utilized electroencephalography (EEG) to monitor the brain activity of students while they wrote essays. The results were telling: those who used ChatGPT to assist with their writing showed significantly less activity in neural networks associated with high-level cognitive processing. Furthermore, these participants struggled to recall the specific contents of their work afterward, suggesting that the AI had performed the “heavy lifting” while the human brain remained relatively disengaged.
The Trust Paradox in the Workplace
This decline in mental engagement isn’t limited to students. Research conducted by Carnegie Mellon University and Microsoft surveyed white-collar professionals to see how AI impacts their daily problem-solving. They found a direct correlation between high confidence in the AI’s ability and a reduction in critical thinking effort.
When workers believe the tool is highly competent, they tend to stop questioning its outputs. This creates a risk of long-term over-reliance, where the user loses the ability to perform independent analysis or verify if a piece of work truly satisfies complex requirements.
A Nuanced View from the Classroom
The impact on younger learners appears to be a double-edged sword. An Oxford University Press (OUP) survey of schoolchildren in the UK found a complex split in perspective:
-
The Benefit: 90% of students felt AI helped them improve at least one skill, such as creative thinking or exam revision.
-
The Drawback: 60% of those same students felt that AI use had a net negative impact on their core schoolwork skills.
-
The “Easy” Factor: A quarter of participants admitted that the technology simply made it too easy to complete assignments without actually understanding the material.
Better Outputs, Poorer Learning?
Academic experts, including those from University College London (UCL), warn of “cognitive atrophy.” This phenomenon has already been observed in the medical field; for example, some radiologists using AI to interpret X-rays performed worse than they did without the tools, likely due to a breakdown in how humans and machines interact.
The fear among educators is that a student might submit a high-scoring, AI-polished essay but finish the course with a shallow understanding of the subject matter. As the saying goes: the output is better, but the learning is worse.
The Path Forward: AI as a Tutor, Not a Ghostwriter
Even developers at OpenAI acknowledge this tension. The company officially discourages “outsourcing” work to ChatGPT. Instead, they advocate for using the AI in a Socratic “study mode”—where the chatbot acts as a 24/7 tutor that breaks down complex components of a question to help the student reach their own conclusion.
Ultimately, the consensus among researchers is that AI should not be viewed as a simple “calculator” for words. To prevent cognitive decline, users must remain informed about how these models reason, stay skeptical of their results, and ensure that the human remains the primary driver of the intellectual process.
About The Author
