Publishers Sue to Obliterate AI Models Trained on Their Work

The Seattle Times is suing Microsoft, a company whose philanthropy arm has long funded its journalism projects.

That stark contradiction underscores the new lawsuit The Seattle Times and New York-based Newsday filed against OpenAI and Microsoft on September 4, 2026, according to TechCrunch. The complaint, lodged in the U.S. District Court for the Southern District of New York, is the latest in a growing pile from publishers. It alleges the tech giants illegally scraped millions of copyrighted articles from behind paywalls to train AI systems like ChatGPT and Microsoft Copilot.

The newspapers describe generative AI as “a snake eating its own tail” in the legal filing, one that could “destroy the very organizations” producing the content it consumes.

“AI products like ChatGPT and CoPilot are touted as producers of content, but in fact they are rapacious consumers, devouring human-authored content and delivering back to the world copies and derivative imitations of that same original content,” the lawsuit states.

They are seeking unspecified damages and a court order to destroy any AI training datasets or models that incorporate their work.

The Lawsuit Accuses AI of Being a Copycat, Not a Creator

The core of the complaint is that the AI models don't just learn from the content, they directly compete with it. The suit alleges the systems can reproduce passages verbatim, closely paraphrase articles, and provide summaries that eliminate the need for readers to visit the publishers’ websites or pay for subscriptions.

This is framed as an existential threat. The Seattle Times disclosed it spends millions of dollars a year to produce its journalism. CEO Alan Fisco, in an email to staff, said the company must defend that investment “from being used without our consent or compensation.”

The legal action is notable not just for its targets, but for the relationship between them. Microsoft is headquartered in the Seattle area and, through its philanthropic arm, has underwritten journalism projects at The Seattle Times. The newspaper also received a grant from a fellowship program funded by a $10 million contribution from Microsoft and OpenAI. This lawsuit signals that prior corporate goodwill does not inoculate against legal action over core intellectual property.

Fair Use Doctrine Faces Its Biggest Test

The defense from OpenAI and Microsoft is expected to be brief and familiar. An OpenAI spokesperson said its models are “trained on publicly available data and grounded in fair use.” Microsoft expressed surprise but said it is “always happy to sit down and explore solutions.”

This sets up the central legal battle: does ingesting copyrighted work to train a commercial AI model constitute fair use?

Publisher's Argument: The use is not transformative in a way that benefits society, but is directly commercial and substitutive. It harms their business by reducing web traffic, the suit cites industry data showing search referral traffic to midsize publishers declined by 47% year-over-year in December 2025. The output can also dilute trademarks by generating fabricated content falsely attributed to the newspapers.

Tech Companies' Argument: The training process is transformative, creating a new, general-purpose tool. The output does not simply regurgitate articles but generates novel responses, and the process is critical for scientific and economic progress. The U.S. Justice Department recently sided with this view in the related New York Times case, filing on September 1, 2026 that a finding of copyright violation would “stifle scientific progress” and hinder American prosperity.

This legal grey zone is why cases are piling up. The New York Times sued in 2023, a group of Alden Global Capital newspapers sued in 2024, and a coalition of 35 publishers sued in 2025. The Seattle Times and Newsday case is another major data point, but no definitive ruling has yet emerged to set a precedent. The outcome will dictate whether AI development becomes a licensed-content industry or remains in its current scrape-and-defend mode.

A Fork in the Road for Media and AI

This lawsuit forces two critical decisions.

First, for other publishers: fight or deal? The path is splitting. Some, like The Associated Press and Axel Springer, have struck licensing agreements with OpenAI. Others are joining the legal fray. The success or failure of these early lawsuits will determine the leverage, and price, in any future negotiations. Every publisher must now calculate whether their content is more valuable as a legal weapon or a licensed asset.

Second, for AI companies: how secure is the training pipeline? A loss in court could mandate the “destruction” of training datasets and models containing unlicensed content. That is a nightmarish technical and financial scenario. It would force a scramble for legally clean data, likely at a much higher cost, potentially slowing development and entrenching advantages for firms with the deepest pockets for licensing. This comes as OpenAI is pushing its technology into new frontiers, including infiltrating U.S. defense with a $1 billion cyber initiative.

XOOMAR Analysis: The timing is pointed. The lawsuit arrived just days after the Justice Department's pro-OpenAI filing in the New York Times case. By filing now, The Seattle Times and Newsday are presenting a counter-narrative directly to the court and the public: this isn't about stifling innovation, but about survival. They are local, privately-owned institutions positioning themselves against what they frame as a coast-to-coast technological threat.

The immediate watch item is whether this case gets consolidated with the New York Times litigation or proceeds separately. More broadly, the pressure is building for a watershed court decision. Until that lands, the relationship between AI and the content it feeds on will remain stuck in a cycle of scraping, suing, and waiting, a costly standoff with the future of both industries in the balance.

Impact Analysis

• It sets a legal precedent for whether AI companies must pay to use copyrighted news content, potentially reshaping their business models.
• The outcome could determine if news publishers can secure revenue from AI training data, helping fund original journalism.
• It challenges the definition of AI as a transformative tool versus a direct competitor that reproduces and replaces original content.