A research team from OpenAI, Duke University, and Harvard University published the first large-scale usage behavior study based on internal ChatGPT data. The study analyzed 26 billion messages from 700 million users from November 2022 to July 2025 through a privacy-preserving automated classification system, revealing ChatGPT usage patterns, demographic changes, and how generative AI creates economic value.
Nov 4, 2025
Research from the University of Texas at Austin, Purdue University and other institutions proposed and verified the LLM Brain Rot hypothesis, and found that continued exposure to spam online text will lead to cognitive decline in large language models, including changes in reasoning ability, long text understanding, security and personality traits.
Oct 31, 2025