Two major news organizations have opened a new legal front in the fight over AI training. The Seattle Times and Newsday sued OpenAI and Microsoft on September 4, 2026, accusing the companies of using published journalism to train and operate AI products without permission or payment.
The case reaches beyond a dispute over licensing fees. It asks a much bigger question: can AI companies ingest news articles, reproduce their words, and compete with the organizations that created them without compensating those publishers?
Publishers Challenge AI’s Use of Their Journalism
The outlets filed a 20-page complaint in federal court in Manhattan. It alleges that OpenAI and Microsoft’s AI models violate their copyrights by ingesting journalism and reproducing it, often verbatim, without compensation.
The complaint argues that generative AI trained on news content threatens to destroy news organizations by competing with them through AI-generated content. That concern arrives as publishers face a steep drop in the traffic that sends readers to their websites: search referral traffic to midsize publishers declined 47 percent in December 2025.
The lawsuit also accuses OpenAI of diluting the outlets’ trademarks. According to the complaint, AI systems generate fabricated content falsely attributed to The Seattle Times and Newsday, creating another risk for publishers whose names depend on accuracy and trust.
The legal demands reach into the systems behind the AI products. The lawsuits seek damages and court orders for the impoundment and/or destruction of datasets and AI models that include the outlets’ articles.
Seattle Times President and CEO Alan Fisco announced the suit on September 4, saying it was about defending content from use without consent or compensation. Fisco said the lawsuit was not about impeding AI innovation, but about protecting the newspaper’s business model.
Newsday spokesperson Tara Rogers said it is imperative to protect the investments made in the organization’s content. The message from both publishers is direct: AI development cannot erase the economic foundation that supports human-created journalism.
OpenAI and Microsoft Defend Their Position
Microsoft said it was surprised by the lawsuit and is open to exploring solutions. OpenAI said its models are trained on publicly available data and grounded in fair use.
That defense places the case inside a much wider legal battle over how copyright law applies to large language models. The US government sided with OpenAI on September 3, 2026, in a lawsuit brought by The New York Times, arguing that training AI models on copyrighted works receives protection under fair use.
The Justice Department said that the “creative possibilities and public benefits that (large language model) training advances far outweigh any competitive harm.” It also argued that The New York Times’ interpretation of the law would create an anti-competitive entry barrier.
In a 20-page document defending AI training, the Justice Department described training AI models on written works as “exceedingly transformative” and said it benefits from fair use. The Trump administration also submitted a 20-page brief in defense of OpenAI’s use of copyrighted material to train large language models.
That brief states that the United States has a strong interest in developing a robust AI industry and retaining global leadership. It references an executive order signed by President Donald Trump and argues that constraining large language model development under a misunderstanding of fair use would hinder progress and economic mobility.
Associate Attorney General Stanley Woodward connected the issue to national priorities, stating, “AI dominance is critical to national security, prosperity and economic mobility.”
A Growing Test for Copyright and AI
The New York Times rejected the government’s position. Graham James, a spokesman for the newspaper, said the Justice Department was siding with “trillion-dollar AI companies” at the expense of American creators.
James also said, “The administration’s proposal to let companies take that content without permission or compensation would undermine the sustainability of the human-created content that a healthy society depends on, and which AI needs to function.”
The dispute follows other major copyright fights. In July 2026, a US judge approved a $1.5 billion settlement between Anthropic and several authors over the use of their works to train AI. Judge William Alsup ordered Anthropic to pay the settlement for using works to train AI, but not for AI training itself, ruling that the use was transformative enough for fair use.
In May 2026, multiple publishers sued Meta for allegedly stealing copyrighted content. Together, these cases show how publishers, authors, technology companies, and the government are testing the boundaries of copyright law while AI systems expand their reach.
The Seattle Times reported on its filing on September 5, 2026, placing the lawsuit at the center of that national argument. One side sees AI training as a transformative process that can unlock creative possibilities, strengthen economic mobility, and preserve US leadership. The other sees unpaid use of journalism as a threat to the business model that produces the content AI needs.
That conflict now moves into court, where decisions about articles, datasets, trademarks, and AI models could shape the next stage of the news industry. The outcome may determine whether future AI products build formal relationships with publishers—or continue fighting over what public availability allows.
Based on
