Microsoft Director Calls AI Scraping 'Largest Theft of Labor in Human History' in NYT Lawsuit Filings

A Microsoft director described the mass scraping of copyrighted work to train artificial intelligence models as 'the largest theft of labor in human history,' according to legal briefs filed in the New York Times' copyright lawsuit against Microsoft and OpenAI. The filings, surfaced this week, also quote OpenAI's own head branding ChatGPT an 'existential threat' to publishers. The documents put the companies' private assessments of their training data practices on the public record, in their own words, inside a case that could reshape how AI firms source the material their models learn from.

The New York Times sued both companies in late 2023, alleging that millions of its articles were used to train OpenAI's models without permission or payment. The case has become a central battleground for the broader fight over whether ingesting copyrighted text counts as fair use. What makes these particular exhibits notable is their source: the companies themselves, not outside critics. Internal language that concedes harm to publishers gives the Times' legal team material it would otherwise have to establish through expert testimony.

The Microsoft director's phrasing is unusually blunt for a document filed in litigation. Calling the practice a theft of labor frames the dispute not as a technical question about fair use but as a labor and compensation issue, which is the argument creative guilds and news organizations have pushed for two years. The OpenAI head's characterization of ChatGPT as an existential threat to publishers cuts the other way for the defense, undercutting any suggestion that the technology poses no meaningful risk to the businesses whose content feeds it.

The revelations arrive as AI companies face a growing stack of infringement claims from authors, musicians, visual artists and news outlets. Some publishers have signed licensing deals with OpenAI, while others, including the Times, chose to litigate. The filings could influence how courts weigh the market-harm prong of fair use, which asks whether a use substitutes for the original work and damages its market. Statements from company leadership acknowledging publisher harm speak directly to that question.

What happens next turns on how Judge Sidney Stein weighs these exhibits and whether the parties reach a settlement before trial. Both companies have defended their training practices as lawful, and Microsoft has not commented publicly on the quoted director's remark. The open question is whether internal acknowledgments of harm become a liability at trial or remain colorful but legally secondary. Either way, the documents give every other plaintiff in the queue a template for what to demand in discovery.