Korva is a scroll of the internet’s best ideas.
What you outsource will atrophy
What outsourcing costs
Most personal uses of AI come down to outsourcing: a task that once took your brain now takes a button press. Outsourcing has always had a cost. Hire a personal chef and you'll probably lose some of your cooking skills. Hire a money manager and you'll probably be less in tune with your finances. Hire an analyst and you'll probably become less proficient with analysis.
Outsourcing thinking itself
That trade is fine for specific, lower-leverage tasks. The worry starts when outsourcing slowly extends into generalized thinking. If more and more of it goes to the models, real knowledge across the board could atrophy, leaving a species of breadth but no depth, defined by surface knowledge.
The MIT study
An MIT study, widely cited though not peer-reviewed, found that using AI tools for writing tasks negatively affected top cognitive function. The small sample and missing peer review mean it proves nothing definitive, but it may be the first of many studies to point in a similar direction.
Has technology weakened a skill before?
A 2020 study by Louisa Dahmani and Véronique Bohbot at McGill University found that people who used GPS more heavily had worse spatial memory when navigating without it. The researchers followed participants over three years, and the decline appeared in those who relied on GPS the most. By contrast, a 2000 study by Eleanor Maguire at University College London found that London taxi drivers, who memorize the city's streets, had a larger posterior hippocampus than other people. The hippocampus is a brain region tied to spatial memory.
Why might easy answers weaken memory?
In a 2011 paper in the journal Science, Betsy Sparrow, Jenny Liu, and Daniel Wegner reported that people remember information less well when they expect to be able to look it up later. They called the pattern the Google effect. Participants instead remembered where to find the information. The brain appears to treat retrievable facts as stored elsewhere and spends less effort encoding them.
What did the MIT study actually measure?
The study came from the MIT Media Lab in 2025 under the title Your Brain on ChatGPT. It divided 54 participants into groups that wrote essays with a chatbot, with a search engine, or with no tools, while researchers recorded their brain activity with EEG. The group using the chatbot showed the weakest connectivity between brain regions. The paper was released as a preprint before peer review.
More from Curiosity Chronicle
