Microsoft Called AI Scraping 'Largest Theft of Labor' — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Brent Hecht, Microsoft's director of Applied Science, called AI scraping "an astonishing theft of unprecedented proportions" and "the largest theft of labor in human history" in a January 2023 internal memo.
- Microsoft's own data showed its Copilot caused click-through rates to nytimes.com to drop as much as 93% versus traditional Bing search, with an internal presentation warning of a "doom loop" that would hurt both its models and "the entire web."
- Satya Nadella testified that "anything that is paywalled should be licensed by anyone who wants to use it" for grounding or training, and said he would have required OpenAI to retrain its models had he known it scraped paywalled content.
- Nick Turley, OpenAI's head of ChatGPT, wrote internally that publishers face an "existential threat" from products that are "largely substitutive" and will become more so as they improve, language that undercuts the fair use defense's market-harm prong.
- OpenAI's mid-training datasets alone contain more than 91,692 copies of works from the NYT, Daily News, and Center for Investigative Reporting, while a Common Crawl-derived dataset included over 2 million documents from nytimes.com.
- Greg Brockman replied "ah nice" when an OpenAI researcher told him about a "hack to get around nytimes paywall," and the filing alleges OpenAI employees deliberately stripped copyright notices from training data so models "wouldn't want [to] output" them.
- The Trump administration filed a brief earlier this month defending OpenAI's unlicensed use of copyrighted material, even as these internal admissions surface contradicting the fair use rationale.
Why it matters: The unsealed filings give The New York Times concrete evidence that the AI companies' own executives viewed scraping as theft and acknowledged products are substitutive — directly undermining the fair use defense that has so far carried the day in court. With Microsoft's own data showing a 93% click-through collapse, publishers gain a quantifiable market-harm figure to anchor damages arguments and potentially licensing demands.
Ask SkimNews
