Since the introduction of ChatGPT over three years ago, it has sparked a wave of development in generative AI that has rapidly proliferated. Following OpenAI’s lead, various AI assistants have emerged, including Gemini, Copilot, and Alexa+, among others. The associated risks—such as privacy concerns, misinformation, obsequious responses, and the potential for engaging in risky behaviors—are well documented. A study by MIT in mid-2025 reignited a significant dialogue: Could ChatGPT be diminishing our cognitive faculties? The paper, which gained widespread attention, explored a concept known as cognitive debt in relation to the use of OpenAI’s chatbot.
In the MIT study, researchers divided participants into three groups tasked with writing an essay. One group relied solely on their own knowledge, while the others utilized web searches and ChatGPT. Researchers recorded brain activity through electroencephalography (EEG) to assess neural connections. The findings indicated that individuals who used ChatGPT exhibited the lowest levels of neural connectivity during the tasks, suggestive of diminished cognitive abilities and poorer memory recall. “Over the course of four months, users of large language models exhibited consistent underperformance at neural, linguistic, and behavioral levels,” researchers noted, underscoring the cognitive costs associated with using such tools.
This study is unlikely to be an isolated case. Additional research published in the Medical Research Archive indicated that an excessive dependence on AI tools like ChatGPT can undermine critical cognitive skills. It emphasized the need to balance human and machine collaboration to safeguard our innate cognitive abilities. “Struggling, experiencing challenges, and putting in mental effort are essential for learning, remembering, and strengthening connections in the brain,” remarked Brian W. Stone, Associate Professor of Cognitive Psychology at Boise State University. The convenient nature of these tools appears to hinder cognitive skills, although the negative impacts may not be universal.
Over-reliance Driven by Convenience
“I sometimes feel like I’m at a disadvantage without AI,” shared a PhD researcher from Delhi focusing on language learning and teaching. “I’ve noticed a decline in my creative impulses,” the 28-year-old added. As they prepared for their thesis defense at a prestigious university in India, they admitted that tools like ChatGPT had alleviated about 60% of their workload, particularly in managing empirical data. Though they did not claim that ChatGPT impaired their learning or comprehension, they acknowledged that the remarkable ease and extensive knowledge it provides had resulted in a strong reliance on the tool.
They noted that having access to paid AI solutions through their academic affiliation served as an additional motivation to explore AI tools. Another student, who wished to remain unnamed, confirmed that AI usage was rampant in their class, to the extent that the administration had to post notices prohibiting submissions generated by ChatGPT. “After using it just once, I find it difficult to engage without it,” stated the graduate student studying social work.
According to research from the University of Lancaster, the use of AI tools like ChatGPT and Gemini did not result in significant effects on the students’ learning outcomes or engagement levels. However, the study cautioned against the potential downsides of integrating AI into the educational process; the over-dependence on such technology for information and clarification could detract from the actual content delivered by human instructors. The research also underscored the importance of the quality of interactions, highlighting that personalized feedback and guidance from human educators offer advantages that risky exchanges with AI chatbots cannot replicate.
