Seattle Times, Newsday sue OpenAI and Microsoft over alleged copyright infringement

Two of America's most established newspapers have filed separate lawsuits against OpenAI and Microsoft, accusing the companies of harvesting their published journalism without permission to train artificial intelligence systems. The Seattle Times and Newsday, both major regional papers with decades of reporting archives, are seeking damages for what they characterize as systematic copyright infringement—the unauthorized use of their original reporting to build commercial AI products.

The legal action marks an escalation in a broader conflict between news organizations and AI developers over the question of who owns the right to use journalistic content for machine learning. As AI companies have raced to build large language models capable of generating human-like text, they have relied on vast quantities of training data scraped from the internet, including millions of articles from newspapers large and small. Publishers argue they never consented to this use and have received no compensation for the value their work provides to these systems.

The timing of these suits reflects growing frustration among media outlets watching their content fuel commercial ventures while their own business models continue to erode. Newsrooms have already faced years of declining print advertising revenue and audience fragmentation. The prospect that their reporting—the product of expensive investigative work, beat reporting, and editorial oversight—could be used to train systems that compete with their own digital products has become a focal point of industry concern.

OpenAI and Microsoft have previously defended their use of publicly available internet content as consistent with fair use doctrine, arguing that training data sourcing is a standard practice in machine learning development. The companies have also suggested that AI systems trained on diverse sources ultimately benefit from exposure to high-quality journalism. Neither company has publicly detailed exactly which news outlets' content appears in their training datasets or in what proportion.

The Seattle Times and Newsday cases will likely hinge on whether courts view the use of copyrighted articles in AI training as transformative—a key test under fair use law—or whether they consider it a direct commercial appropriation of intellectual property. The newspapers will need to demonstrate both that their content was used without permission and that this use caused them measurable harm, whether through lost licensing revenue, diminished competitive advantage, or other economic injury.

These lawsuits arrive as other media organizations, including The New York Times, have also begun exploring legal remedies or negotiating licensing agreements with AI companies. Some publishers have attempted to reach commercial arrangements with OpenAI and other developers, seeking payment for the use of their archives. Others have implemented technical barriers to prevent their content from being scraped for training purposes. The legal landscape remains unsettled, with no clear precedent yet established for how courts will treat journalistic content in the context of AI model development.

The outcome of the Seattle Times and Newsday cases could reshape how AI companies source training data and whether they must obtain explicit permission or negotiate licensing fees with publishers. It may also influence whether other news organizations pursue similar litigation or whether the industry moves toward standardized compensation models. For now, the question of who bears the cost of training the next generation of AI systems—and whether that cost should be borne by the companies building them or absorbed by the creators of the content they use—remains unresolved.