I would like to add that the authors of the study provided a FAQ that says the following:
Is it safe to say that LLMs are, in essence, making us "dumber"?
No. Please do not use the words like “stupid”, “dumb”, “brainrot”, "harm", "damage", and so on. It does a huge disservice to this work, as we did not use this vocabulary in the paper, especially if you are a journalist reporting on it.
Additional vocabulary to avoid using when talking about the paper
In addition to the vocabulary from Question 1 in this FAQ - please avoid using "brain scans", "LLMs make you stop thinking", "impact negatively", "brain damage", "terrifying findings".
And I would like to point out that the twitter OP uses at least two of these terms, the Time story quoted in the middle of this post discusses "harms" that the study authors are not certain are harmful and that the tags are full of people saying that this is terrifying, horrifying, will rot people's brains, we're cooked, we told you so, etc, which are specifically things that the authors didn't want to have as the takeaway for the story.
I'd also like to note that the OP of the twitter thread is the COO of a company that sells AI assistants whose entire twitter presence is using AI-written threads about AI to generate social media traffic to drive to his AI sales company.
His takeaway from all of these, and the original thread on this post, is that people should use the AI *he* sells.
The MIT preprint is an interesting preliminary study that has suggestive findings about the way that habitual AI use changes users' behavior; it also had interesting points to make about how the search engine arm of the study was noticeably influenced by terms that had been promoted by google ads, and it recorded that people who used search engines had brain connectivity that was reduced compared to people writing only from their minds. But nobody is talking about the horrifying, harmful cognitive debt making us dumber caused by search engines (and they shouldn't be!)
The guy who is calling this cognitive atrophy is an AI shill who wants to use terms that the study's authors don't stand behind in order to make AI seem more powerful than it actually is. The study (and the Time article screencapped above) both note that the highest level of mental engagement was from people who had previously written essays with no assistance writing a final essay *with* AI assistance (so in that circumstance the use of AI strengthened cognitive response). To quote that article directly: "The second group, in contrast, performed well, exhibiting a significant increase in brain connectivity across all EEG frequency bands. This gives rise to the hope that AI, if used properly, could enhance learning as opposed to diminishing it."
So, basically what we've done here is let a 200-word AI summary of a 200+ page paper take the burden of cognitive load off our shoulders instead of taking the time to think critically about where we're getting our information from.
I continue to be less concerned about AI rotting our brains than I am about how infrequently people click through links on social media.
I'd recommend reading some of the study if you've got time. I found the ways that the English teachers and the AI used by the study differed in their analysis of the papers - because, please note, the authors of the study used LLMs for some of the data - really interesting.
kaledo-art
stormborns
yousharknotpass
dededos
jacobvanloon
izzy806
gummygunk
