The copyright fight between The New York Times, other news publishers, Microsoft and OpenAI has turned on a deceptively simple question: can AI companies train products on journalism without permission and call it fair use?

In unsealed filings released Thursday, the publishers argued that the companies’ own records make that defense harder to sustain. They cited a January 2023 memo from Microsoft’s director of applied science, Brent Hecht, which described mass AI scraping as “an astonishing theft of unprecedented proportions” and possibly “the largest theft of labor in human history.”¹

The plaintiffs say the warnings were not confined to abstract copyright concerns. A Microsoft presentation from January 2024 allegedly forecast a traffic “doom loop”: AI answers would drain publishers of readers, undermining the web content that AI systems themselves need. In the filing’s account, Copilot’s answer engine cut click-through rates to the Times domain by as much as 93% compared with traditional Bing search.²

OpenAI’s internal discussion, as characterized by the publishers, pointed in the same direction. ChatGPT chief Nick Turley reportedly called the threat to publishers “existential,” saying chatbots were “largely substitutive” and would become more so as they improved. The filings also allege employees discussed a way around the Times paywall, to which OpenAI president Greg Brockman responded, “ah nice.”³

Publishers contend that evidence of substitution, extensive copying and alleged paywall evasion shows the firms knew their products could displace the market for original reporting. They argue that the stakes reach beyond one lawsuit: without payment or traffic, fewer journalists will produce the reliable work AI tools depend on.

Microsoft rejects the attempt to turn internal remarks into an admission. A spokesperson said its products are transformative fair use rather than substitutes for news sites, and stressed that Hecht’s documents “reflect one employee’s individual perspective, are not a legal analysis, and do not represent the company’s views.”³ The company says Nadella’s deposition remarks described broad shifts in how people consume information, not conclusions about the copyright claims before the court.