Internal documents from Microsoft and OpenAI, unsealed in a lawsuit led by The New York Times, reveal that employees internally characterized the scraping of news content for AI training as a form of theft. Microsoft Director of Applied Science Brent Hecht described the practice as the largest theft of labor in human history, while OpenAI staff acknowledged that commercial products trained on news pose an existential threat to publishers.
Data cited in the court motion indicates that Microsoft recorded click-through rate declines of 83–93 percent for some news sites and 51–94 percent for others. Internal communications show that OpenAI executives and engineers were aware that chatbots function as substitutes for news platforms, noting that users have little incentive to click on original links when information is provided directly within the AI interface.
The documents also detail instances where OpenAI staff discussed bypassing paywalls, with President Greg Brockman responding positively to a report about a hack to access NYT content. Additionally, the motion alleges that OpenAI used a third-party dataset containing 1.8 million articles from The New York Times, despite internal awareness that the data was restricted from commercial use.
Microsoft is accused of selling a dataset originally purchased for Bing to OpenAI for model training without consulting the news organizations. Plaintiffs argue that these actions constitute a doom loop where AI companies destroy their own content supply chains. They are seeking a court ruling that such scraping does not qualify as fair use to ensure all AI firms operate on an even footing.
