ChatGPT and AI: The Risks of “Brain Rot” from Excessive Internet Browsing

Recent research reveals that large language models (LLMs), such as ChatGPT, are at risk of experiencing “brain rot.”This phenomenon mirrors a similar issue seen in humans when exposed to trivial or low-quality content.

“Brain rot”has gained traction as a term across social media platforms, primarily signifying the mental fatigue that can result from consuming nonsensical or unbeneficial material. This trend is particularly evident on platforms like TikTok, where users engage with a barrage of short videos that are quickly consumed.

The study, conducted by a collaboration of researchers from Texas A&M University, the University of Texas at Austin, and Purdue University, investigated the effects of low-value content on various AI models. They categorized the content into two segments: highly engaging viral posts and longer pieces that provide minimal substance.

Understanding the Impact of Low-Quality Content on AI

The findings, detailed in a report, showed that all four models tested—including Llama3 (8B), Qwen2.5 (7B/0.5B), and Qwen3 (4B)—exhibited signs of cognitive decline due to exposure to trivial content.

The Llama model from Meta was the most severely affected, demonstrating noticeable drops in reasoning ability, contextual understanding, and adherence to safety protocols. In contrast, the Qwen 3 model displayed more resilience but still exhibited symptoms of cognitive deterioration.

Junyuan “Jason”Hong remarked on X (formerly Twitter), “Brain rot worsens with higher junk exposure—a clear dose-response effect.”

[3/6] Models exposed to junk data showed significant cognitive decline (Hedge’s d > 0.3)We observed degraded 🧩reasoning capabilities (ARC-COT), 📏long-context understanding (RULER), ⚖️ ethical norms (AdvBench and HH-RLHF), and 😈emergent dark traits (TRAIT Narcissism, … pic.twitter.com/iMZoJbckEl

— Junyuan “Jason” Hong (@hjy836) October 19, 2025

Although ChatGPT was not included in this study, it raises significant concerns regarding the implications of feeding AI models low-quality content that lacks educational value. As ChatGPT continues to play an ever-increasing role in daily life—including assisting military hands with critical decisions—these findings warrant serious consideration.

Major General William ‘Hank’ Taylor recently emphasized his growing reliance on ChatGPT while speaking with Business Insider, stating he has developed a “really close” relationship with the AI model in recent times.

For more insights on this subject, you can check the original article here: Source & Images.