Internal documents from Microsoft and OpenAI, released on September 17, reveal internal assessments regarding the impact of AI tools like ChatGPT and Copilot on the news industry. Brent Hecht, Microsoft's head of applied sciences, described the scraping of news content for AI training as a massive theft of labor. He also criticized the application of the fair use doctrine in this context.
Nick Turley, head of ChatGPT, characterized AI products trained on news as an existential threat to publishers. Microsoft documents outlined a doom loop scenario where declining publisher revenue leads to a lack of quality content, ultimately degrading both AI models and the internet. Microsoft data showed click-through rate declines for some publishers ranging from 83% to 93%, and for others from 51% to 94%.
CEO Satya Nadella admitted under oath that chatbots replace news platforms by providing information directly on the AI interface instead of directing users to the original source. Evidence presented by publishers included an OpenAI employee reporting a hack to bypass The New York Times paywall, which President Greg Brockman acknowledged. Plaintiffs also alleged that Microsoft sold Bing data to OpenAI without publisher consent.
Lawyers for news organizations demonstrated that chatbots could reproduce verbatim fragments of copyrighted text when asked to evaluate article bias. A Microsoft spokesperson stated that the internal comments reflected the views of individual employees rather than the company's official position. Conversely, attorney Steven Lieberman noted that these documents expose the companies' internal awareness regarding their practices.
