Unsealed documents in the New York Times' lawsuit against OpenAI and Microsoft reveal the companies' own internal warnings that their practices would start a "doom loop" damaging the web. Their documentation described scraping data for AI training as the "largest theft of labor in human history" and