Microsoft Director Privately Admitted AI Was the “Largest Theft of Labor in Human History,” Unsealed Court Documents Show

In a stunning revelation that sends shockwaves through the burgeoning artificial intelligence industry, recently unsealed and unredacted court documents in the landmark copyright infringement lawsuit brought by the New York Times against OpenAI and Microsoft expose candid admissions from top AI executives. These private statements, now public under the threat of perjury, characterize the foundational data acquisition practices of AI development as nothing short of “an astonishing theft of unprecedented proportions” and even “the largest theft of labor in human history.” This development profoundly escalates the stakes in an already contentious legal battle, offering the NYT powerful ammunition against the tech giants.

The unredacted motion, filed on behalf of the New York Times, meticulously leverages the words of key figures within the AI ecosystem to underscore the plaintiff’s core arguments. Among the most damning quotes is one attributed to Brent Hecht, a director of applied science at Microsoft. Hecht is quoted describing the expansive data collection and training process for AI models as the “largest theft of labor in human history.” This striking characterization directly aligns with the NYT‘s central accusation: that OpenAI and Microsoft have systematically copied “millions of [the] Plaintiffs’ copyrighted articles in their entirety without permission to produce substitutive commercial AI products.” Such an admission from within the defendants’ own ranks casts a long shadow over their claims of legitimate data use and fair practice.

Further solidifying the NYT‘s position are the statements from OpenAI cofounder and president Greg Brockman. Brockman is quoted acknowledging that “publishers” face an “existential threat” from generative AI, adding that these AI systems “are largely substitutive, period, [and] will get more and more substitutive as they get better.” This candid assessment from a leading figure at OpenAI, the creator of ChatGPT, lends significant credence to the New York Times‘ argument that AI models are designed not merely to assist, but to directly compete with and ultimately replace human-generated content, thereby undermining the economic model of news organizations. The notion of “substitution” is critical to the NYT‘s case, as it directly addresses the potential market harm, a key factor in copyright infringement analyses.

These extraordinary admissions, made under circumstances demanding truthfulness, are presented by the plaintiff’s counsel as effectively “eviscerat[ing]” any viable claim to fair use—a crucial legal defense often invoked by AI companies. Fair use, a doctrine under U.S. copyright law, permits limited use of copyrighted material without acquiring permission from the rights holder, typically for purposes such as criticism, comment, news reporting, teaching, scholarship, or research. The determination of fair use traditionally hinges on four factors: (1) the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes; (2) the nature of the copyrighted work; (3) the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and (4) the effect of the use upon the potential market for or value of the copyrighted work. The NYT argues that the sheer volume of material ingested, the commercial intent, and the “substitutive” nature of the AI output directly violate the spirit and letter of fair use principles. Tellingly, Microsoft’s own Brent Hecht evidently concurs with this assessment, recognizing that for the defendants to prevail on a fair use defense “would arguably ‘make a complete mockery of the idea of ‘fair use.”” This internal skepticism regarding their own legal defense strategies is a powerful point for the prosecution.

The lawsuit itself, filed in late 2023, is a pivotal moment in the ongoing debate surrounding AI and intellectual property. The New York Times accused OpenAI and Microsoft not only of illegally vacuuming up its extensive archive of original news content, but, crucially, of using this stolen intellectual property to train AI products designed to generate content that directly competes with, and often reproduces, the NYT‘s journalism. The complaint detailed instances where OpenAI’s ChatGPT and Microsoft’s Copilot allegedly generated verbatim excerpts from NYT articles, sometimes presenting inaccurate or fabricated information, yet still attributing it to the newspaper, thereby eroding trust and devaluing the original content.

The implications of these admissions extend far beyond the immediate confines of this lawsuit. They highlight a fundamental tension at the heart of the AI revolution: the rapid advancement of technology built upon a vast, often uncompensated, appropriation of human creativity and labor. Hecht’s phrase, “largest theft of labor in human history,” resonates deeply with concerns raised by artists, authors, musicians, programmers, and other creators globally. These groups increasingly find their life’s work ingested into AI models without consent, credit, or remuneration, only to see AI-generated outputs threaten their livelihoods and the very concept of original authorship. The economic impact on creative industries, particularly journalism, is a central concern. If AI models can produce articles that mimic the style and content of established news outlets, the incentive for readers to subscribe or visit original sources diminishes, directly impacting advertising revenue and subscription models that sustain independent journalism.

This case is also a critical test for the future of copyright law in the digital age. Existing legal frameworks, developed long before the advent of large language models, are struggling to keep pace with the technological paradigm shift. The outcome of the NYT v. OpenAI/Microsoft lawsuit could set a significant precedent, influencing how AI developers acquire and utilize training data, potentially leading to new licensing models, compensation structures for content creators, or even regulatory interventions. Governments and legislative bodies worldwide are closely watching, grappling with how to balance innovation with the protection of intellectual property rights.

As the legal proceedings enter the “middle stretch,” the presiding judge is now tasked with evaluating the merits of the arguments presented, including these newly revealed admissions. A key decision point will be whether the case is allowed to proceed to a full trial, which Axios recently reported could happen by sometime in 2027. Should the case advance, it promises to be one of the most closely watched legal battles in recent memory, with profound consequences for both the future of AI development and the sustainability of creative industries, particularly news and journalism. The direct testimony from within the AI companies themselves, now made public, has undeniably shifted the narrative, placing the onus squarely on the defendants to convincingly counter their own executives’ characterizations of their data practices.

The gravity of the situation cannot be overstated. These admissions, dragged into the light by the strictures of legal discovery, paint a stark picture of an industry aware of the potentially illicit foundations of its technological marvels. The battle between the New York Times and the AI titans is more than just a copyright dispute; it is a fundamental clash over the value of human creativity, the ethics of technological progress, and the very future of how information is generated, consumed, and compensated in an AI-driven world.