The Seattle Times and Newsday sued OpenAI and Microsoft on Friday, September 4, 2026, alleging in federal court that the tech companies copied their journalism without permission to train and operate their AI products, according to Reuters.
The lawsuit adds two major American newspapers to the fast-growing list of media companies taking legal action against the makers of generative AI systems. For more context on this story, see our ongoing AI news.
What the Newspapers Allege
According to the complaint, OpenAI and Microsoft copied articles from both outlets to train and run their AI systems. The Seattle Times Co. — the parent company of Microsoft's hometown daily — joined forces with New York's Newsday, which broke the news of its own suit on Friday.
GeekWire reported that the Seattle Times is suing Microsoft and OpenAI for allegedly training their AI on its journalism, a notable escalation given the newspaper operates in Microsoft's backyard.
The newspapers are seeking an unusually aggressive remedy: the destruction of their copyrighted works and of any AI systems or datasets containing their material, Reuters reported. They argue that AI-generated answers built on their reporting reduce visits to their websites and undercut subscriptions — the revenue that funds the journalism in the first place.
Why This Lawsuit Matters
The Remedy Sought Is the Real Story
Most earlier copyright suits against AI companies have demanded damages. This one goes further, asking the court to order the destruction of datasets and models that contain the plaintiffs' work. If courts ever granted such relief, the operational impact on an AI company would be severe — far beyond a financial settlement.
A Local Paper Taking On Its Hometown Giant
The optics are striking. The Seattle Times has covered Microsoft for decades as the region's dominant tech employer. Now the paper's parent company accuses the Redmond company of building billion-dollar products on the back of its reporting without a license.
Newsday, a Pulitzer-winning daily on Long Island, brings a similar claim, arguing that chatbots that summarize or restate its articles keep readers from ever reaching its own pages.
Part of a Much Bigger Legal Wave
The case lands amid intensifying legal pressure on AI developers over training data. Record labels including Sony and Warner have sued Anthropic over song lyrics, a German court handed music-industry group GEMA a landmark win over Suno, and multiple authors, visual artists, and outlets have pending claims against OpenAI and Microsoft.
For publishers, the core complaint is consistent across these cases: AI products can generate answers that substitute for the original journalism, diverting traffic while the underlying content was used without compensation or consent.
OpenAI has consistently argued in prior cases that training on publicly available content is fair use. Microsoft, as OpenAI's largest backer and cloud provider, has been named alongside it in most of the major media suits.
What Happens Next
The case will take months, if not years, to move through the federal courts. In the meantime, AI companies are racing to sign licensing deals with publishers, and courts are being asked for the first time to weigh whether mass copying of news archives to train models counts as fair use or infringement.
For local and regional newspapers fighting for survival, the stakes could not be clearer — their reporting is showing up in chatbot answers, whether or not anyone paid for it.
What Each Side Is Likely to Argue
OpenAI and Microsoft have consistently defended their training practices in earlier cases by arguing that ingesting publicly available web content to build AI systems is transformative fair use. Expect the same framework here, along with the point that chatbot answers are not substitutes for full articles in every case.
The newspapers, by contrast, will point to the direct overlap: an AI answer that restates a Seattle Times investigation or a Newsday scoop can satisfy a reader who never clicks through, converting what was once a subscription-driving visit into a lost one. The requested remedy — destroying the datasets and any models trained on them — is designed to force the court to take that economic argument seriously rather than treat infringement as a cost of doing business.
Legal analysts have long noted that destruction remedies are rarely granted in copyright cases, but the ask matters: it raises the settlement value of the case and adds pressure on AI companies to license news content rather than litigate the question for years.
The Broader Picture for Local News
The suit also highlights an asymmetry in the AI economy. Building frontier models requires billions in compute, but the raw material — decades of reporting, archives, and local accountability journalism — was produced on shoestring budgets by newsrooms now watching their work power products they never agreed to support.
Whatever the outcome, the Seattle Times and Newsday case is likely to become one of the defining tests of whether American courts will extend copyright protection to the AI training era — or leave publishers to negotiate, article by article, with the companies whose products can now imitate them.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
