Timnit Gebru, co-lead of Google's Ethical AI team, disclosed on Twitter that she was out of the company. The trigger was Google's demand that she retract, or remove Google-affiliated authors from, a co-authored paper on the risks of large language models, "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?" The paper pointed to the environmental cost of ever-larger models, the amplification of bias present in training data, and the fact that a model appearing to understand meaning may only be arranging words by probability. Gebru said she tried to negotiate conditions, and that an email treating her as having resigned went to her reports immediately afterward. The two sides give conflicting accounts of how she left, and thousands of people, including more than 2,000 AI researchers, signed letters of protest. The episode confronted the industry with the question of how an AI company handles internal research that criticizes its own technology.