Newly unsealed filings in The New York Times' three-year-old copyright suit against Microsoft and OpenAI show a Microsoft executive describing the companies' AI training practices in terms that undercut their own legal defense. Microsoft Director of Applied Science Brent Hecht wrote that mass scraping of news content for AI training was "an astonishing theft of unprecedented proportions," and possibly "the largest theft of labor in human history," according to unredacted material cited in the plaintiffs' motion for summary judgment Microsoft exec called AI scraping the “largest theft of labor in human history” - Ars… (arstechnica.com). Hecht separately wrote that the scraping plan made "a complete mockery of the idea of 'fair use'" — the doctrine Microsoft and OpenAI have relied on in court Microsoft exec called AI scraping the “largest theft of labor in human history” - Ars… (arstechnica.com).
The filings also attach numbers to the harm the plaintiffs say that scraping caused. Microsoft's own data showed click-through rates to The New York Times' domain falling as much as 93% when users got Copilot-generated answers instead of traditional Bing search results, and separate figures cited from the filing put the range at 51% to 94% across news plaintiffs Microsoft exec called AI scraping ‘the largest theft of labor in human history,' new… (techcrunch.com)Microsoft exec called AI scraping the “largest theft of labor in human history” - Ars… (arstechnica.com). Microsoft has responded that Hecht's comments reflect one employee's personal view, not the company's legal position Internal Microsoft warning brands AI scraping the 'largest theft of labor in human… (tweaktown.com).
What the filings allege Microsoft and OpenAI actually did
The unsealed material describes specific conduct, not just training methodology. It alleges the companies bypassed paywalls "undetected," built training datasets through mass scraping, and stripped copyright notices from the data before use Microsoft exec called AI scraping ‘the largest theft of labor in human history,' new… (techcrunch.com). In one exchange OpenAI engineer Nick Ryder told OpenAI President Greg Brockman that crawlers had found "a hack" to get around the Times' paywall; Brockman replied, "Ah, nice" Microsoft exec called AI scraping the “largest theft of labor in human history” - Ars… (arstechnica.com). Fair use permits unlicensed use of copyrighted work in specific circumstances such as parody or news reporting, and one factor courts weigh is whether the use substitutes for the original in its market Microsoft exec called AI scraping ‘the largest theft of labor in human history,' new… (techcrunch.com). The plaintiffs argue the click-through data does exactly that — Hecht's own internal document called the effect a "doom loop" that would "hurt the performance of our models and the entire web at the same time," and warned it was "highly unusual that an end-product threatens the economic foundations of its essential suppliers" Microsoft exec called AI scraping ‘the largest theft of labor in human history,' new… (techcrunch.com)Microsoft exec called AI scraping the “largest theft of labor in human history” - Ars… (arstechnica.com). OpenAI's head of ChatGPT, Nick Turley, wrote separately that products like ChatGPT posed an "existential threat" to publishers because they were "largely substitutive" and would become more so as they improved Microsoft exec called AI scraping ‘the largest theft of labor in human history,' new… (techcrunch.com).
Where this sits against a business still compounding
None of this shows up in Microsoft's own numbers yet. Fourth-quarter fiscal 2026 revenue was $90.0 billion, up 18% year over year, with Microsoft Cloud revenue at $59.3 billion, up 27%, and Azure surpassing $100 billion in annual revenue for the first time EX-99.1 - SEC.gov. Microsoft 365 Copilot passed 30 million paid seats EX-99.1 - SEC.gov. Analysts at Stifel estimated in July 2026 that about half of Azure's fiscal 2026 revenue growth came from OpenAI alone Microsoft to disclose Azure revenue as part of segment changes (cnbc.com) — the same commercial relationship now central to the copyright case. Microsoft's stock traded around $493 in mid-September 2026, down from a 52-week high of $553.72, with no indication in that pricing that scraping litigation is being treated as a distinct risk factor MSFT: Microsoft Corporation - Stock Price, Quote and News - CNBC. Government procurement has kept expanding in parallel: the General Services Administration's OneGov agreement, announced in September 2025, projects up to $3.1 billion in first-year federal savings on Microsoft 365, Copilot and Azure Multi-Billion Dollar GSA OneGov Agreement with Microsoft Brings Steep Discounts for…, and Microsoft's government cloud offering carries 99 separate federal authorizations under FedRAMP Microsoft 365 Government Community Cloud & Supporting Services (fedramp.gov).
The gap between the internal language and the external numbers is itself the open question in this case, not something either side has resolved. Microsoft's position — that one executive's assessment is not the company's legal position — has not been tested against a jury, and the plaintiffs say they are prepared to go to trial on "compelling evidence of substitution" Internal Microsoft warning brands AI scraping the 'largest theft of labor in human… (tweaktown.com)Microsoft exec called AI scraping the “largest theft of labor in human history” - Ars… (arstechnica.com). Whether that evidence moves a jury the way it has moved headlines is what the trial, not the filing, will decide.