OpenAI sued for training on more than 10 million news articles; NYT material was heavily represented
Court papers say OpenAI used more than 10 million articles — nearly a third from The New York Times — and publishers seek damages.
OpenAI is facing a major copyright lawsuit after plaintiffs say the company trained its models on more than 10 million news articles, with nearly a third of that material coming from The New York Times, according to court filings published on September 18, 2026.
The complaint reproduces internal descriptions of the data collection. Microsoft’s applied science director Brent Hecht is quoted in the filing as calling the acquisition "a shocking theft on an unprecedented scale," and even "perhaps the largest theft of human labor in history." Microsoft later told reporters that Hecht's remarks reflect "the view of one employee speaking personally" and do not represent the company's position.
Claims, counterclaims and the involvement of the US government
The New York Times originally filed suit against OpenAI and Microsoft three years ago in federal court in New York, accusing the companies of unlawfully taking and using copyrighted content to train OpenAI's flagship model, ChatGPT. Microsoft first invested in OpenAI in 2019. Other publishers have since joined the suit, including Ziff Davis, the parent company of CNET, the owner of Mother Jones, The Intercept and several local newspapers.
Plaintiffs are seeking damages for each article they say was taken and used without permission, though the total potential award has not been determined. The publishers asked the court for a ruling in their favor without proceeding to a full trial; if U.S. District Judge Sidney Stein grants that request, the case would not go to jury trial, but a final decision is not expected before 2027.
OpenAI has since reached licensing agreements with numerous media companies after ChatGPT’s launch at the end of 2022, but publishers and critics warn that generative AI can siphon traffic away from news sites and undermine journalism’s business model. Developers have tried to mitigate those concerns by including citations and links in model responses; an OpenAI engineer told investigators that "No matter how visibly we show links, users do not want to click them."
At the start of September, the U.S. Department of Justice filed a brief supporting OpenAI and Microsoft, arguing that the technology advances "scientific progress," promotes "economic growth" and serves "national security." Both OpenAI and Microsoft contend their use of news material is transformative and protected by the U.S. doctrine of fair use.
Photo: press material from the event
