While the headline is catchy, this post is not as well thought-out as my previous thoughts on the relationship between AI assistants and psychology. It is more of an aside to a recently published study.
MIT’s Media Lab has conducted a study on how AI assistants affect our thinking in terms of intellectual output and brain activity:
Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing
Abstract: With today’s wide adoption of LLM products like ChatGPT from OpenAI, humans and businesses engage and use LLMs on a daily basis. Like any other tool, it carries its own set of advantages and limitations. This study focuses on finding out the cognitive cost of using an LLM in the educational context of writing an essay. […]
The use of LLM had a measurable impact on participants, and while the benefits were initially apparent, as we demonstrated over the course of 4 months, the LLM group’s participants performed worse than their counterparts in the Brain-only group at all levels: neural, linguistic, scoring.Preprint on arXiv: https://arxiv.org/pdf/2506.08872
The results confirm something I have observed over the past month in several situations: The ease of asking ChatGPT or Claude a question about any simple or difficult problem and simply accepting the result as true leads to worse discussions. People cannot answer follow-up questions and struggle to truly understand what ChatGPT has provided them (answers, code, explanations, etc.). Previously, I have argued that ChatGPT makes us lazy thinkers because we no longer solve hard problems ourselves. Of course, this is a bit exaggerated, because it really depends on how you use these tools in particular situations.
The study aligns closely with this observation, and I would be very surprised if replication studies did not yield similar results across different manipulations and outcomes.
Connecting the dots between social media use (old hat, I know) and LLM use, I see the danger that LLMs will only worsen the issues we already face with social media and smartphone usage: declining attention spans, a lack of rigorous thought in everyday contexts, poorer communication skills, and more. I even suspect that the same groups most negatively affected by social media use will be the same ones most impacted by the effects of AI assistants in both their personal and professional lives.