The dispute dates to late 2023, when The New York Times sued OpenAI and Microsoft over the alleged use of its reporting to train generative AI. Eleven other publishers later joined the case, and Judge Sidney H. Stein is now weighing summary-judgment motions as documents are gradually unsealed.¹

Those filings have handed news organizations unusually sharp ammunition. A January 2023 memo attributed to Microsoft Director of Applied Science Brent Hecht described the mass collection of online work as “the largest theft of labor in human history,” while another warning called it “an astonishing theft of unprecedented proportions.”²

The publishers’ broader case is that the companies saw the commercial danger early. An internal Microsoft presentation reportedly warned of a “doom loop”: AI answer products could pull readers from publishers, weaken the web’s content supply, and ultimately hurt the models built on it. The filing cited Microsoft data showing Copilot’s answer engine cut click-through rates to the Times domain by as much as 93% against conventional Bing search.²

OpenAI’s internal discussion, as described in the filings, points in the same direction. Nick Turley, the head of ChatGPT, wrote that publishers faced an “existential threat” from products that were “largely substitutive” and would become more so as they improved. The filings also allege that OpenAI staff discussed a way around the Times paywall, to which President Greg Brockman replied, “ah nice.”³

Microsoft contests the interpretation. Its spokesperson said Hecht’s comments “reflect one employee’s individual perspective, are not a legal analysis, and do not represent the company’s views.” The company maintains that its AI products are transformative fair use rather than replacements for news sites.³

That is the fault line now before the court: publishers see insiders acknowledging an extraction-and-substitution model; Microsoft sees disputed internal debate, not proof that its AI business broke copyright law.