A group of researchers from Texas A&M, the University of Texas at Austin and Purdue have discovered that large language models (LLMs) are just as susceptible to “brain rot” as humans. The study, which fed viral X posts to the LLMs, discovered that the short-form “junk” content or “low-quality (sensationalist) content” caused them to suffer “cognitive decline.”

Wow, you don’t say.

Their conclusion was that exposure to junk content reduced the models’ reasoning skills, impaired reading comprehension, diminished ethical norms and resulted in "emergent socially undesirable personalities.”

This might be the most disturbing AI paper of 2025 ☠️

Scientists just proved that large language models can literally rot their own brains the same way humans get brain rot from scrolling junk content online.

They fed models months of viral Twitter data short, high-engagement… pic.twitter.com/ARGbLeGYld

Advertisement

— Alex Prompter (@alex_prompter) October 20, 2025

In layman’s terms, after spending too much time scrolling through the slop-feeds of X, these models became dumber, meaner and developed new unsociable personalities.

That sounds about right.