Brain rot content is giving AI poorer information judgment, just like humans. A new study from three U.S. universities found that large language models fed a diet of viral but low-quality social media content experience cognitive decline—reduced reasoning abilities, degraded memory, and even becoming “more psychopathic” according to ethical benchmarks. Researchers fed Meta’s Llama and Alibaba’s Qwen models social posts optimized for engagement—think sensational text or overhyped script, the kind that goes viral with young viewers—and found the damage couldn’t be easily undone through retraining. The findings mirror research on humans, which shows this kind of content has a detrimental effect on cognitive abilities, so much so that “brain rot” was Oxford Dictionary’s word of the year in 2024. For Gen Z, who coined the term to describe their own doomscrolling habits, it’s darkly poetic: the AI trained on their feeds is now suffering the same fate. (Wired)
👀 Read more from YPulse: Tweens Are Talking Brainrot—They Name Their Top 10 Slang Terms
