Seattle Times and Newsday sue OpenAI and Microsoft

In 2024, Microsoft and OpenAI co-funded a $10 million AI journalism fellowship through the Lenfest Institute that included both The Seattle Times and Newsday as participants^[1]. Now, in September 2026, those same two newspapers are suing their former philanthropic partners in the U.S. District Court for the Southern District of New York, accusing them of methodically scraping paywalled journalism to train and operate ChatGPT, Microsoft Copilot, and Bing's AI features^[2]. The two defendants' reactions underline the awkwardness of that history: Microsoft's spokesperson said the company was surprised by the lawsuit but remained willing to "sit down and explore solutions to this type of dispute"^[3], a conciliatory tone that contrasts with OpenAI's more defensive line that its models "train on publicly available data grounded in fair use, helping hundreds of millions improve daily lives"^[4], without addressing the specific paywall-bypass allegations. Seattle Times CEO Alan Fisco has tried to frame the suit not as hostility toward AI but as defense of costly reporting, saying the paper must protect content it spends "millions of dollars a year to produce" from uncompensated use^[3].

The complaint frames generative AI products as consumers rather than producers of content, arguing AI systems are "rapacious consumers, devouring human-authored content" in a dynamic the newspapers liken to "a snake eating its own tail"^[5]- training practices that, if left unchecked, erode the very journalism the models depend on. As evidence, the 38-page filing cites an 88-word passage allegedly lifted from the Seattle Times' Pulitzer Prize-winning Boeing 737 MAX coverage, and alleges hundreds of thousands of articles were used without authorization^[1][4]. Rather than seeking damages alone, the plaintiffs are asking the court to order impoundment or destruction of copies of their copyrighted works, along with any training datasets or AI models that incorporate the material^[4]- a remedy that, if granted, would carry far more technical disruption for OpenAI's and Microsoft's products than a monetary settlement.

This suit does not stand alone. The New York Times filed the precedent-setting copyright case against OpenAI and Microsoft in December 2023 over the use of millions of articles to train ChatGPT, a case that remains ongoing^[2][5]; CNN separately sued AI search company Perplexity in 2024 over similar scraping claims^[2]. Anthropic already settled comparable claims from authors and publishers for $1.5 billion in 2025^[4], and roughly 400 local newspapers filed their own coalition lawsuit against AI companies in June 2026^[1][6]. Regulatory pressure is building in parallel: a German court ruled in July 2025 that works stored within an AI model's parameters constitute infringement at the point of storage, and the EU's General-Purpose AI Code of Practice copyright chapter took effect in August 2025, with OpenAI as a signatory committing not to circumvent paywalls going forward - though that commitment does not cover the conduct alleged here, which predates it^[6].

The newspapers' economic argument rests on a specific data point: search referral traffic to midsize publishers fell 47% year-over-year in December 2025, which the complaint attributes to AI-generated answers reducing readers' need to visit news sites or pay for subscriptions^[4]. That argument sits in direct tension with a stated federal position: the U.S. Department of Justice argued on September 1, 2026 that AI development serves the national interest and warned that copyright-violation findings against AI developers could "stifle scientific progress"^[4]. Early public reaction to the filing has been muted rather than heated - coverage on X has been largely a neutral wire-service relay of the news from outlets like Reuters and Newsmax, with no visible pushback yet from media-law or AI-policy voices, consistent with how fresh the story still is.