The escalating legal war over generative AI has a new front: The Seattle Times and Long Island daily Newsday have filed a federal lawsuit accusing OpenAI and Microsoft of copying their journalism to train AI systems without permission.
The complaint, brought in the U.S. District Court for the Southern District of New York, alleges that the companies scraped hundreds of thousands of articles from the papers’ sites — including content locked behind paywalls — to power products such as ChatGPT, Microsoft Copilot, and Bing’s AI-powered search features.
According to the filings, those AI tools can reproduce long passages from Seattle Times and Newsday stories and closely paraphrase others, giving users the substance of their reporting without sending traffic back to the newspapers’ own pages.
The publishers argue that this behavior amounts to direct and vicarious copyright infringement and that the systems erode subscription and advertising revenue by replacing visits to their sites with AI-generated answers.
Seattle Times Co. and Newsday are seeking unspecified monetary damages along with court orders to impound or destroy copies of their copyrighted works, any training datasets built from them, and AI models that incorporate that material.
The lawsuit also asserts trademark claims, accusing OpenAI of diluting their brands when its systems hallucinate and falsely attribute fabricated content to the Seattle Times and Newsday, potentially confusing readers about what the papers actually published.
This case lands amid a fast-growing wave of media lawsuits against generative AI companies, including a landmark action filed by The New York Times in late 2023 that similarly accuses OpenAI and Microsoft of using millions of Times articles without authorization to train ChatGPT and related models.
In March 2026, reference publishers such as Merriam-Webster and Encyclopedia Britannica launched their own “massive” copyright claims against OpenAI over the ingestion of dictionary and reference content into AI training corpora, further widening the front of the fight.
Other media and publishing outfits have joined in: WikiHow has sued over training on its how-to articles, while Penguin Random House has targeted OpenAI over alleged copying tied to a popular German children’s book series, all against the backdrop of a recent Justice Department brief urging a Manhattan judge to find that OpenAI’s use of news articles for training is lawful.
Seattle Times and Newsday’s complaint leans hard into the stakes for local journalism, warning that generative AI could leave the news industry “broken beyond repair” as search referrals to mid-size outlets drop sharply and readers increasingly get summarized answers instead of clicking through to original stories.
The suit notes that nearly 400 local newspapers have recently filed their own case against OpenAI and Microsoft, and highlights the irony that the tech firms have helped fund some of The Seattle Times’ reporting initiatives even as they allegedly exploit its content at industrial scale.
The timing is striking: the filing arrives just as OpenAI begins rolling out GPT-6 Astra, its most advanced model to date, and as AI assistants like Copilot become increasingly woven into everyday devices and apps.
For tech and media fans, the outcome will help determine how much training data future models can legally ingest, whether AI companies must negotiate broad licenses for news archives, and how tools like Copilot and ChatGPT present and attribute the journalism they draw upon.
If courts side with publishers and order the destruction of models trained on unlicensed news content, AI developers could be forced to rebuild their systems with cleaner, more expensive data—or pushed into sweeping licensing deals that more fairly compensate the outlets whose work fuels the bots.
