The rapid integration of Artificial Intelligence (AI) into our daily and professional routines is sparking serious concerns among scientists. New research indicates that heavy dependence on tools like ChatGPT could be weakening essential human cognitive abilities, including critical thinking and problem-solving skills.
Brain Scans Reveal Reduced Cognitive Activity
A pivotal study released earlier this year by the Massachusetts Institute of Technology (MIT) provided neurological evidence of this worrying trend. Researchers monitored the brain activity of 54 participants from MIT and nearby universities using Electroencephalography (EEG), which involves placing electrodes on the scalp.
The findings, reported by the BBC, were clear: individuals who used ChatGPT to write essays showed significantly less activity in brain networks linked to deep cognitive processing during the task. Furthermore, these participants struggled later to explain the logic of their essays or quote from them, unlike those who wrote without AI assistance.
The MIT team concluded that their work highlights "the pressing matter of exploring a possible decrease in learning skills" due to AI outsourcing.
The "Intellectual Autopilot" Trap in the Workplace
Separate research, a joint effort by Carnegie Mellon University and Microsoft (which operates Copilot), reinforces these fears in a professional context. Their study surveyed 319 white-collar workers who used AI tools weekly, analysing over 900 task examples from data analysis to compliance checks.
The key discovery was a direct link between high confidence in an AI tool and a drop in "critical thinking effort." When workers trust the AI's output, they often stop verifying it, entering a state researchers termed "intellectual autopilot."
The study warned that while generative AI can boost efficiency, it may inhibit critical engagement with work, leading to long-term overreliance and diminished independent problem-solving skill.
Echoes from Classrooms and Clinics
The concern spans different sectors. In the United Kingdom, an Oxford University Press (OUP) survey in October found that six in 10 students believed AI had negatively impacted their academic skills.
While OpenAI's ChatGPT—boasting over 800 million weekly users—has published prompts to help students, experts like Prof Wayne Holmes from University College London (UCL) argue this is insufficient. He told the BBC that "no independent evidence at scale" exists for the tools' effectiveness or safety in education, urging more research before widespread student encouragement.
Even in high-stakes fields like medicine, the effects are mixed. A 2025 Harvard Medical School study found AI assistance improved some clinicians' performance but damaged others' for unclear reasons, prompting calls for more research into human-AI interaction.
The collective message from global institutions is unambiguous: as AI becomes ubiquitous, understanding and mitigating its potential to erode fundamental human cognitive skills is an urgent priority.
