OpenAI and Microsoft employees privately acknowledged their artificial intelligence products compete with the news outlets whose articles trained them, an unsealed court filing alleges.
OpenAI’s head of ChatGPT Nick Turley wrote that the company’s products “are largely substitutive, period” and will become even more so as the technology improves, according to a brief The New York Times and other news publishers made public Thursday in their Manhattan copyright case against the tech companies. Microsoft Director of Applied Science Brent Hecht warned that “millions of people around the world will soon consider large models ‘hoovering up’ all their work to be an astonishing theft of unprecedented proportions,” the brief stated in part.
The Department of Justice previously told a federal judge that AI training is fair use because models learn patterns from text rather than duplicate it, an argument the companies’ internal messages seemingly undercut.
Turley also wrote that publishers face an “existential threat” from those products, according to the brief.
The brief also cites a document describing the practice as potentially the “largest theft of labor in human history,” and says Hecht recognized that a fair use win for the companies would arguably “make a complete mockery of the idea of ‘fair use.’”
Microsoft and OpenAI each did not immediately respond to the DCNF’s request for comment.
A tech policy expert said the statements cut against the companies’ legal defense.
“These companies admit in their own internal communications to things that completely contradict their legal defense and their theory of fair use,” Evan Swarztrauber, principal at CorePoint Strategies and a former adviser to Federal Communications Commission (FCC) Chairs Ajit Pai and Brendan Carr, told the DCNF. “[O]ne of the key tests for fair use is whether original product that was copied is going to be competed against and substituted for, and they literally say … [i]t’s substitution.’”
Microsoft’s own documents mentioned similar concerns about AI’s effect on publishers.
“Our AI content strategy has started a ‘doom loop’ that will hurt the performance of our models and the entire web at the same time,” a Microsoft document quoted in the filing read in part. “It is highly unusual that an end-product threatens the economic foundations of its essential suppliers, but that is the situation we have created for our LLM business.”
Another internal document cited in the brief describes large language models as “a product that destroys its supply chain.”
Microsoft CEO Satya Nadella testified that talking to chatbots “has substituted” for visiting original sources and that “anything that is paywalled should be licensed” for training, the brief stated. Nadella also said he would have mandated OpenAI to retrain models if he knew it trained on paywalled material.
OpenAI co-founder and President Greg Brockman wrote that the models are “particularly good at predicting text of news articles,” according to the filing.
When employee Nick Ryder told Brockman about “a hack to get around nytimes paywall” to help “scrape” the Times site, Brockman replied “ah nice,” the brief stated. OpenAI’s corporate representative testified he did not know of “any effort to detect or remove paywalled content” from training data.
Microsoft spokesman Alex Haurek told The New York Times that Hecht’s writings reflected “one employee’s individual perspective,” were not legal analysis and did not represent Microsoft’s views.
OpenAI argued that training on publicly available material is fair use in Jan. 2024.
“You can’t be expected to have a successful AI program when every single article, book or anything else that you’ve read or studied, you’re supposed to pay for,” President Donald Trump said in July 2025 when he announced the AI Action Plan, Politico reported.
Swarztrauber argued the case could push the industry toward a market for licensing publishers’ work.
“[T]he reason this market has been stifled is because the official position of the AI industry has been: we don’t have to. We owe you nothing,” Swarztrauber said. “There’s no reason that the most sophisticated companies in human history cannot figure out a way to pay people for their work.”
(DCNF)
