OpenAI and Microsoft were aware their work would trigger a 'doom loop' for the internet

Newly unsealed documents from the New York Times' copyright lawsuit against OpenAI and Microsoft contain sharply critical assessments written by the companies themselves. Internal materials described the rush to scrape web data for AI training as setting off a "doom loop" that could harm the internet, called the practice "the largest theft of labor in human history," and said it amounted to "a complete mockery of the idea of fair use." The filings give the Times fresh ammunition in a case that could shape the legal foundations of the entire generative AI industry.

Hecht's internal critiques take center stage

Many of the most striking statements in the filings trace back to Brent Hecht, Microsoft's Director of Applied Science. His written assessments, now part of the public court record, include the sweeping characterization of AI training data collection as historically unprecedented theft of human creative work, alongside pointed skepticism about whether fair use doctrine could plausibly cover the companies' practices.

The "doom loop" warning suggests internal concern that aggressive AI-driven content scraping could degrade the open web itself — if publishers respond to having their work consumed by models by locking down or removing content, the quality and availability of training data and web information generally could deteriorate over time.

Microsoft moves to contain the damage

Microsoft has sought to separate itself from Hecht's statements. Company spokesperson Alex Haurek told The Verge that the comments referenced in the filings do not represent the company's position, signaling that Redmond intends to treat the internal writings as personal opinions rather than admissions of corporate wrongdoing.

That distinction could matter enormously in court. Plaintiffs in copyright cases often seize on internal documents to argue that defendants knew their conduct was legally risky. Microsoft's response indicates the defense strategy will hinge on framing such statements as individual views that never reflected official policy or decision-making.

Why the case matters for the AI industry

The New York Times sued OpenAI and Microsoft in December 2023, alleging that the companies used millions of its articles without permission to train and power products like ChatGPT. The lawsuit has become one of the most consequential legal tests of whether training AI models on copyrighted material qualifies as fair use — a question with implications far beyond journalism.

The unsealed documents arrive as courts and regulators worldwide weigh similar claims. Other authors, news organizations, and rights holders have filed comparable suits, all watching for signals about how judges will treat the argument that AI training is transformative.

What to watch next

Both companies will likely continue battling over which internal communications can be admitted as evidence and how they are characterized. As discovery proceeds and potential rulings on fair use approach, statements like Hecht's — and the disputes over their meaning — will remain a recurring flashpoint in the case.