Taking comfort in this. Apparently, consuming social media has the same effect on AI chatbots as human beings: brain rot.

In a paper in Nature, researchers at UT Austin put a group of AI chatbots on a high social media diet for an extended period of time. The bots—powered by multiple LLMs, including Llama3, Qwen, and DeepSeek—significantly slowed down, worsened, or did not follow any reasoning. These models have different designs: Qwen, OpenAI’ o1, and DeepSeek R1 are reasoning models that produce reasoning steps to arrive at an answer. Llama is an instruction-tuned language model with less advanced reasoning ability.

The effect of social media content on all the models was similar: they became slow in the head and showed poor or no reasoning.

Sounds familiar?