Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task

shouldn’t just accept change unquestionably. > Of course, this all depends on using AI to enhance cognition and access to knowledge, as opposed to just letting a machine write all your code for you without review, Yegge-style. And the latter is exactly what is going to happen and is already happening in large enough quantity that it’s going to be a serious problem. > But the one about "endangering human development" is wholly in our individual hands. You can use AI to help you learn, or to replace the need to learn. That completely ignores the loss of skill that happens without you realising, as you lean more on a tool. https://www.thelancet.com/journals/langas/article/PIIS2468-1... https://arxiv.org/abs/2506.08872 This is nothing new. We already know that e.g. heavy GPS use makes us weaker at navigating on our own. https://www.nature.com/articles/s41598-020-62877-0 > One real lesson from this is perhaps that we need to teach people how to use AI in ways that benefit their development, not just their output. Yes, that is a good goal. But good luck achieving it.

style. https://arxiv.org/abs/2409.01754 https://arxiv.org/abs/2508.01491 https://aclanthology.org/2025.acl-short.47/ https://arxiv.org/abs/2506.06166 https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing https://osf.io/preprints/psyarxiv/wzveh_v1 https://arxiv.org/abs/2506.08872 https://aclanthology.org/2025.findings-acl.987/ https://aclanthology.org/2025.coling-main.426/ https://aclanthology.org/2025.iwsds-1.37/ https://www.medrxiv.org/content/10.1101/2024.05.14.24307373v... https://journals.sagepub.com/doi/full/10.1177/21522715251379... https://arxiv.org/abs/2506.21817

We actually need to socialize and be around people to remain healthy. You miss out on all of these benefits with chat bots. And there's growing evidence that you don't learn as much when using them [0]. Consider when using them is appropriate and maybe don't rely on them for everything. [0] https://arxiv.org/abs/2506.08872

of writing an essay. We chose essay writing as it is a cognitively complex task that engages multiple mental processes while being used as a common tool in schools and in standardized tests of a student's skills. Essay writing places significant demands on working memory, requiring simultaneous management of multiple cognitive processes. A person writing an essay must juggle both macro-level tasks (organizing ideas, structuring arguments), and micro-level tasks (word choice, grammar, syntax). In order to evaluate cognitive engagement and cognitive load as well as to better understand the brain activations when performing a task of essay writing, we used Electroencephalography (EEG) to measure brain signals of the participants. In addition to using an LLM, we also want to understand and compare the brain activations when performing the same task using classic Internet search and when no tools (neither LLM nor search) are available to the user. [0] https://arxiv.org/pdf/2506.08872

Entities Recognition (NERs), n-grams, ontology of topics within each group. EEG analysis presented robust evidence that LLM, Search Engine and Brain-only groups had significantly different neural connectivity patterns, reflecting divergent cognitive strategies. Brain connectivity systematically scaled down with the amount of external support... And then: "...in this study we demonstrate the pressing matter of a likely decrease in learning skills based on the results of our study." I'm not sure "likely decrease in learning skills" is quite right here. [0] https://arxiv.org/pdf/2506.08872v1

brains and offloads much of our cognitive activity. When working with models that require a high level of abstraction and optimization, offloading these tasks to LLMs that essentially scrape and combine code from various sources can lead to disaster. For something like a basic CRUD app such factors are not nearly as important. I can see the future convergence of two clear trends in software: No-code tools and AI tools. The development of AI agents which handle not only the generation of code but the full development cycle of creating, testing, and deploying projects with a limited level of programming knowledge required. As an additional note: Both of these software trends aim to automate and replace many responsibilities of the programmer in a commercial sense. I believe that future programmers will need to become familiar with these emerging technologies in order to remain competitive in the job market. Namely, math skills. Linear regression, multivariable calculus, the sort of math applied by scientists and researchers. If you do not understand the mathematics at the core of neural networks, now is the time to learn.

have the DR required equipment on-site to ensure generators are up right after Hurricane Sandy passes, and pre-emptively handle issues before they become outright disasters. Anyone thinking you can have helpdesk do this isn't right in the head. > I think LLMs have absolutely nothing to do with this. There are a growing number of places that get lock-in contracts where the SA related services they offer are being done by helpdesk staff manning LLM based chatbots without disclosure, blindly. This obviously works right up until it doesn't, and by the time people notice the fundamental issue, the competent people the company needs won't be around anymore with a vicious race to the bottom. These helpdesk people may be manning 3 or more chatbots simultaneously in a foreign country while pulling in the same as retail fast food in California. No SA can compete with nominally free slave labor, and these people don't care about Information Governance and unauthorized data disclosures. Any seasoned SA worth their salt will be capable of automation in python and powershell at a fairly advanced level if they're MS centric, and python/bash for Linux. DevOps Engineer, Architect, SRE all have quite a lot of overlap with a competent SAs, but SAs aren't getting these jobs. There's no path to getting those jobs, and the actual stats and hiring being done seems to indicate there aren't many people getting these positions, and they are almost universally grey beards. No one going in, no one coming out, what happens when they die/age out/flame out... The related IT systems fail in growingly chaotic trends, and related production grinds to a halt. It becomes just like the ending of Ayn Rand's Atlas Shrugged, but without the new beginning. LLMs aren't a substitute for people, but companies are using them as such, and there is an increasing body of literature showing that LLM use increases cognitive debt. (https://arxiv.org/abs/2506.08872)