AI
Microsoft's own director warned AI scraping was "largest theft of labor in human history"
Unsealed court documents reveal internal Microsoft warnings that training AI on news content posed existential risks to publishers and contradicted fair-use claims.
Ars Technica
Microsoft and OpenAI have long kept sensitive materials confidential during ongoing copyright disputes with news outlets over unauthorized use of journalistic content for AI training. But internal documents that were recently unsealed in litigation paint a stark picture of how the tech giants privately understood the stakes of their strategy.
In a motion for summary judgment unveiled Thursday, news organizations led by The New York Times disclosed internal memos suggesting that Microsoft and OpenAI's own leaders recognized the severity of their scraping practices before rolling out products like ChatGPT and Copilot.
A Microsoft Director's Stark Assessment#
Among the most damaging revelations is a series of warnings from Brent Hecht, Microsoft's Director of Applied Science. In internal correspondence, Hecht characterized the practice of harvesting news articles for AI model training as "an astonishing theft of unprecedented proportions" — and went further still, labeling it perhaps the "largest theft of labor in human history." The documents also show Hecht challenging Microsoft and OpenAI's public defense that such use constitutes fair use, noting instead that their scraping strategy made "a complete mockery of the idea of 'fair use.'"
Existential Concerns Acknowledged Internally#
At OpenAI, the gravity of the situation was similarly recognized. Nick Turley, who leads ChatGPT, described in an internal message the potential for commercial AI systems trained on news to pose an "existential threat" to traditional news publishers—essentially enabling AI products to displace human journalists and news organizations as information sources.
One Microsoft document articulated what it called a "doom loop": a self-reinforcing cycle in which AI training on declining-quality web content would simultaneously degrade model performance while harming the publishers whose material sustained the web ecosystem. The memo observed that "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 with respect to its 'content supply chain.'"
These newly surfaced communications stand in sharp contrast to the public positions both companies have maintained throughout the dispute.

