NYT Lawsuit Reveals Microsoft Exec Calls AI Scraping ‘Largest Theft of Labor’ and OpenAI Calls ChatGPT an ‘Existential Threat’ to Publishers

Bottom line

The New York Times’ legal brief shows that a Microsoft director described AI‑driven web scraping as "the largest theft of labor in human history" and that OpenAI’s head of ChatGPT called the model an "existential threat" to publishers, underscoring the high‑stakes copyright fight over the data used to train large language models (LLMs).

Legal backdrop

  • The New York Times sued OpenAI and Microsoft in late 2023 for alleged copyright infringement stemming from the use of thousands of its articles to train ChatGPT and Copilot.
  • In September 2026 the Times filed a motion for summary judgment, attaching internal memos from both companies that had been sealed or redacted at the defendants’ request.
  • The brief is the first public source confirming the companies’ own internal language about the controversy.

“The brief cited an internal memo dated January 2023 by Microsoft director of Applied Science Brent Hecht, where he allegedly said, ‘Millions of people around the world will soon consider large models “hoovering up” all their work to be an astonishing theft of unprecedented proportions’ and also called it ‘the largest theft of labor in human history.’” – Tom’s Hardware summary of the NYT filing.

Microsoft’s internal assessment

  • Brent Hecht’s memo frames AI scraping as a systemic labor‑theft problem, warning that the practice harms the “content supply chain” that fuels LLMs.
  • The memo adds that "almost no one intended for content they created to be used in this fashion, nor are they compensated for its use."
  • Hecht also described a “doom loop” where reduced click‑through rates (e.g., a 93 % drop for NYT when users rely on Copilot) diminish publisher revenue while simultaneously degrading model performance because the web’s content ecosystem erodes.

“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.’” – quoted in the NYT brief.

OpenAI’s internal warning

  • Nick Turley, head of ChatGPT, labeled the chatbot an "existential threat" to publishers, noting that the model is "largely substitutive" and will become "more substitutive as it gets better."
  • An OpenAI engineer testified that "no matter how prominently we show the links, users won’t click," implying that the model’s answers effectively replace the need to visit the original source.
  • Another researcher, Nick Ryder, discussed a “hack to get around NYT paywall,” to which OpenAI president Greg Brockman replied, "ah nice," suggesting awareness of deliberate circumvention tactics.

“OpenAI Head of ChatGPT Nick Turley said in internal communications that the AI chatbot is an ‘existential threat’ to publishers as they are ‘largely substitutive’ and ‘will get more and more substitutive as they get better.’” – Tom’s Hardware.

Fair‑use defense under pressure

  • AI firms have so far relied on the U.S. copyright doctrine of fair use (17 U.S.C. § 107) to justify large‑scale web scraping for model training.
  • A 2024 court upheld Anthropic’s use of publicly available books as fair use, but the NYT brief shows that Microsoft and OpenAI executives explicitly recognized the market‑harm their scraping caused.
  • The purpose and effect factors of fair‑use analysis are now complicated by internal admissions that the models substitute for the original content and deprive publishers of traffic and revenue.

“Microsoft CEO Satya Nadella said in a deposition that ‘anything that is paywalled should be licensed by anyone who wants to use it…for grounding or training’ and that if he ‘had been made aware that OpenAI has scraped and trained on information that was behind a paywall,’ the company would have required OpenAI ‘to retrain its models.’” – Tom’s Hardware.

Publisher impact and industry reaction

  • The NYT claims Copilot reduced its click‑through rate by up to 93 % compared with Bing search, directly hurting ad revenue.
  • Industry observers fear a cascade effect: as LLMs become more capable, users will increasingly bypass original articles, accelerating the doom loop described by Microsoft.
  • The lawsuit could set a precedent for how paywalled or copyrighted content is treated in AI training pipelines, potentially forcing AI firms to obtain licenses or to exclude such material entirely.

Community perspectives (Hacker News comments)

  • ttquantia reported that Google’s AI Overview feature reproduces entire passages from news articles, echoing concerns that search‑engine AI also infringes on copyright.
  • adamddev1 described a personal experience where an LLM reproduced a novel term the commenter invented, without attribution, calling the behavior "egregious" and "plagiarism."
  • muchdoubt warned that Microsoft could secretly harvest Copilot data to gain a competitive edge, suggesting the risk extends beyond public scraping.
  • fxtentacle expressed skepticism about Microsoft’s claim that it would have forced OpenAI to retrain models if it had known about paywall scraping, implying the statement may be a legal posture rather than a genuine commitment.
  • Several commenters (e.g., bitwize, es​chaton) highlighted the broader ethical dimension, comparing AI‑driven data extraction to historic forms of labor exploitation.

“Boy, it's a good thing IP as we know it is going the way of deodand laws! Otherwise the AI companies might be in real trouble!” – @bitwize.

What this means for the AI ecosystem

  1. Legal risk escalates – The NYT’s internal‑memo evidence could undermine the fair use shield that AI developers have relied on.
  2. Business models may shift – If courts side with publishers, AI firms will need to negotiate licensing deals or develop clean training datasets, increasing costs.
  3. Transparency pressure rises – Regulators and the public are demanding clearer disclosure of what data fuels LLMs; internal memos like Hecht’s provide a rare glimpse into corporate awareness.
  4. Potential market realignment – Publishers might monetize their content through AI‑specific licensing, creating new revenue streams while protecting their editorial labor.

Takeaway

The NYT lawsuit’s newly disclosed internal memos reveal that top executives at Microsoft and OpenAI acknowledge AI scraping as a massive appropriation of human labor and a direct threat to publishers’ business models, exposing a stark contradiction between their public fair‑use arguments and private recognition of market harm. The outcome of this case could reshape the legal and economic foundations of generative AI.

Sources

Related