# OpenAI and Microsoft Staff Privately Called Their Own AI an Existential Threat to News > Source: > Published: 2026-09-22 13:48:01+00:00 *Unsealed court filings show OpenAI's head of ChatGPT and a Microsoft director privately admitted their AI products threaten the survival of the news publishers suing them.* OpenAI and Microsoft executives privately admitted their AI products pose an existential threat to the news industry they scrape for training data. That's according to unsealed court filings. Nick Turley, OpenAI's head of ChatGPT, wrote internally that publishers face an existential threat from AI products that are already largely substitutive, period and will only get more and more substitutive as they get better. The line surfaced in a joint brief filed by The New York Times and other publishers suing OpenAI and Microsoft for copyright infringement, unsealed this past week. It's the kind of admission plaintiffs usually can only dream of finding. Microsoft's director of applied science, Brent Hecht, went further. He called the underlying copying an astonishing theft of unprecedented proportions and, in a separate note cited in the filing, possibly the largest theft of labor in human history. Hecht also flagged what he described as an accidental cover-up: the companies' own struggles to identify how much Times, Daily News and Center for Investigative Reporting content had ended up baked into their models. Here's the number that makes his warning concrete. OpenAI's mid-training datasets contain more than 91,692 copies of works published by the Times, the Daily News and the Center for Investigative Reporting, according to the filing. The company scraped over 10 million articles total. Nearly a third came from the Times alone. [OpenAI Shuts Down Sora's API This Week, Ending Its Billion Dollar Disney Deal](https://startupfortune.com/openai-shuts-down-soras-api-this-week-ending-its-billion-dollar-disney-deal/) OpenAI's Sora video API shuts down for good on September 24, 2026, just days after the news broke. The shutdown, which reportedly saved OpenAI roughly $1 million a day in compute costs, also killed a $1 billion Disney licensing deal that let Sora users generate clips with Marvel, Pixar, and Star Wars characters. - [why is OpenAI shutting down Sora API](https://startupfortune.com/openai-shuts-down-soras-api-this-week-ending-its-billion-dollar-disney-deal/) - [what happened to OpenAI Sora video generator](https://startupfortune.com/openai-shuts-down-soras-api-this-week-ending-its-billion-dollar-disney-deal/) Then there's the paywall exchange. When OpenAI researcher Nick Ryder told cofounder and president Greg Brockman about a hack to get around the Times' paywall for scraping, Brockman wrote back: ah nice. Two words, but they matter. Digiday reported that publishers are now arguing this kind of paywall workaround eviscerates OpenAI's fair-use defense, since fair use has historically leaned on the idea that copying happened from freely accessible material, not paid content deliberately circumvented. ## From Inference to Admission Publishers have spent two years arguing that generative AI competes directly with the journalism it was trained on. Proving that in court usually means building an inference from circumstantial evidence. Traffic drops. AI Overviews. Chatbot answers that summarize a story without sending anyone to read it. What the Times and its co-plaintiffs found instead was their opponents' own staff saying it plainly, in writing, before the lawsuit forced anyone to choose sides. Microsoft's response has been to distance itself from Hecht. A company spokesperson told reporters that his statements reflect one employee's individual perspective and do not represent the company's views. That's a standard corporate move. It doesn't erase the document. Turley's comments are harder to wave off, since they describe the core business reality he was hired to manage: how well ChatGPT replaces the products it was trained on. None of this settles the case. Fair use is still the central legal question, and OpenAI and Microsoft haven't conceded that training on copyrighted news articles is illegal, only that insiders worried, at various points, about how it would look and what it would cost the publishers who supply the raw material. But admissions like these shift the weight of a trial. A jury doesn't need to infer substitution from a chart of declining pageviews when a company's own head of product wrote existential threat in a message that was never meant to become a court exhibit. ## The Irony Under Hecht's Warning He wasn't just describing harm to a rival industry. He was describing a supply chain problem for AI itself. Chatbots depend on real-time news and quality reporting to stay useful and current, and Hecht called it highly unusual for a product to threaten the economic survival of the very suppliers it needs. If the news industry it trains on shrinks or shuts down, AI loses the raw material that keeps its answers accurate. That's not a hypothetical risk analysts are speculating about. It's a warning that came from inside Microsoft's own applied science team. Trial in the NYT v. OpenAI and Microsoft case hasn't been scheduled yet. But the publishers now have exhibits that read less like allegations and more like a confession. **Also read:** [Nscale Files for a $35 Billion IPO That Tests the Neocloud Story](https://startupfortune.com/nscale-files-for-a-35-billion-ipo-that-tests-the-neocloud-story/) • [Dario Amodei will ask the UN Security Council to slow down the AI race this week](https://startupfortune.com/dario-amodei-will-ask-the-un-security-council-to-slow-down-the-ai-race-this-week/) • [Inspur Used Its Aivres Shell to Send $5.6 Billion in Nvidia Chips to China](https://startupfortune.com/inspur-used-its-aivres-shell-to-send-56-billion-in-nvidia-chips-to-china/) [Newsom Signs Adam's Law, California's Toughest AI Chatbot Safety Rules](https://startupfortune.com/newsom-signs-adams-law-californias-toughest-ai-chatbot-safety-rules/) Gov. Gavin Newsom signed SB 1119, known as Adam's Law, on September 10, 2026, making California the first state to require AI companion chatbots to detect suicidal ideation in minors, notify parents, undergo independent safety audits, and face lawsuits when they fail. The law is named for Adam Raine, a 16-year-old whose 2025 death led to a... - [california AI chatbot safety law for minors](https://startupfortune.com/newsom-signs-adams-law-californias-toughest-ai-chatbot-safety-rules/) - [adam's law requires AI detection of suicidal ideation](https://startupfortune.com/newsom-signs-adams-law-californias-toughest-ai-chatbot-safety-rules/) *This article is posted in [AI News](https://startupfortune.com/category/ai/), check it out for more related stories.* ## Join the discussion [Open in the community →](https://startupfortune.com/community/) Almost there. Sign in and your reply posts straight away.