Microsoft AI scraping comments surface in New York Times copyright case
An unsealed plaintiffs’ brief attributes the “largest theft of labor” remark to Microsoft’s Brent Hecht; Microsoft says it was not company policy.
By Renata Fuchs · Policy Reporter
· 3 min read
Microsoft AI scraping theft comments have entered the copyright fight between The New York Times, OpenAI and Microsoft after an unsealed brief from news plaintiffs attributed a stark internal assessment to Brent Hecht, Microsoft’s director of applied science. According to the plaintiffs’ filing, Hecht described scraping news for AI training as perhaps the “largest theft of labor in human history,” while Microsoft said the remarks reflected one employee’s view rather than the company’s position.
The material surfaced in a summary-judgment motion in the Times-led litigation, which was filed in 2023. The publishers allege that OpenAI and Microsoft copied their journalism without permission to train AI systems and that chatbot outputs can compete with the original reporting. Those allegations have not been established by a court in the material available.
The reported language is consequential less as a legal conclusion than as evidence the plaintiffs hope will undercut the defendants’ fair-use argument. Fair use is the copyright doctrine that can permit unlicensed use in certain circumstances; Microsoft maintains its AI uses are transformative fair use. Whether that defense applies here remains unresolved.
What did Microsoft say about the AI scraping comments?
Microsoft told Ars Technica that Hecht’s documents expressed “one employee’s individual perspective,” were not legal analysis and did not represent the company’s views. A spokesperson also said Microsoft’s court filings explain why its uses comply with copyright law and why Copilot is not a substitute for publishers’ journalism.
That distinction matters. An employee’s reported private assessment does not amount to an admission by Microsoft, nor does the unsealed brief constitute a judicial finding that training practices were unlawful.
What the plaintiffs say the internal material shows
According to Ars Technica’s account of the brief, Hecht also characterized broad news scraping as an “astonishing theft of unprecedented proportions” and said the plan would make a “complete mockery” of fair use. The plaintiffs also cited internal statements that they say recognize how AI answers could reduce visits to publishers’ sites and damage the news businesses that produce the underlying work.
Those claims are central to the publishers’ theory that the AI products are substitutive rather than transformative. The plaintiffs have also alleged that the companies’ systems generated extensive verbatim overlap with articles. Microsoft disputes the substitution premise; OpenAI did not immediately respond to Ars Technica’s request for comment.
What remains unknown
The reporting comes with an important evidentiary limit. TechCrunch reported that much of the newly visible material appears in the Times’ brief, while the underlying exhibits remain sealed. The quoted snippets therefore lack their original public context.
For operators building products on web-scale data, the episode illustrates the gap between public fair-use defenses and internal concern about the economic effect of AI answers on content suppliers. The court has not decided the copyright and fair-use questions raised in the case. For now, the record supports a narrower conclusion: plaintiffs have put a Microsoft researcher’s reported criticism before the court, and Microsoft has disavowed it as corporate policy.
This story draws on original reporting from Ars Technica.
