Leaked Files Reveal Microsoft and OpenAI Executives Warned AI Threatens News Publishers

Unsealed court documents show that executives at Microsoft and OpenAI privately warned that training AI models on news content poses existential threats to publishers and significantly reduces web traffic. Internal records revealed that Microsoft's Brent Hecht questioned fair-use defenses, while OpenAI's Nick Turley noted chatbots directly replace the need for users to visit news websites. These internal admissions weaken the legal defense by AI companies that training on news content is purely transformative fair use with no market harm to creators. It also highlights a potential 'doom loop' where AI starves publishers of web traffic and revenue, ultimately degrading the quality of reliable news data available for future AI training. Microsoft internal data indicated traffic drops between 51% and 94% for certain news organizations involved in copyright litigation, with CEO Satya Nadella testifying that chatbots take clicks away from original source sites. Additionally, documents revealed Microsoft created a filtering mechanism that potentially made it harder for copyright holders to identify whether their content was used for AI training.