Seattle Times and Newsday Sue OpenAI and Microsoft Over AI Training

The Seattle Times and Newsday filed a joint lawsuit in the U.S. District Court for the Southern District of New York on Friday, accusing OpenAI and Microsoft of unauthorized use of copyrighted journalism to train generative AI models [1, 2]. The litigation marks a significant escalation in the media industry's effort to protect business models, as publishers claim AI-generated content reduces traffic and subscription revenue by providing users with alternatives to original reporting [1, 4].

The core of the dispute

The lawsuit alleges that OpenAI and Microsoft "methodically scraped" news articles, including content protected by paywalls, to develop products such as ChatGPT, Microsoft Copilot, and AI-powered Bing features [1, 3, 4]. According to the publishers, these AI models can reproduce entire passages or closely paraphrase original reporting, which they argue dilutes their trademarks and creates false information by attributing fabricated content to the newspapers [1, 3, 4].

The publishers cite industry data indicating that search referral traffic to midsize publishers declined by 47% year over year in December 2025, attributing this loss in part to the rise of AI-generated search alternatives [2]. Seattle Times CEO Alan Fisco stated that the organization spends millions of dollars annually to produce its content and must defend it from being used without consent or compensation [2, 4]. In response, an OpenAI spokesperson maintained that its models are trained on publicly available data and are grounded in the "fair use" doctrine [2, 4]. Microsoft, which has invested billions into OpenAI since 2019, expressed surprise at the filing but indicated a willingness to discuss potential solutions [2, 3].

Legal and competitive landscape

The case arrives amid a broader legal pushback from the media sector. The New York Times initiated a similar copyright battle against the same defendants in 2023, and other publishers, including a group of 35 newspapers representing over 400 community outlets, have previously filed suit [1, 2, 3]. While some organizations like the Associated Press and Vox Media have opted to partner with OpenAI to license their material, others are seeking damages and the destruction of AI training sets that incorporated their work [1, 3].

The federal government has recently signaled a stance that complicates the publishers' position. In a filing this week, the U.S. Department of Justice argued that AI development is a matter of national interest, suggesting that a ruling against the tech companies could stifle scientific progress and hinder economic mobility [2, 3]. The government further contended that the potential benefits of AI models "far outweigh any competitive harm" that might befall individual publishers [2].

What to watch

• Judicial Precedent: Whether the court accepts the "fair use" defense for AI training or mandates the destruction of existing datasets, as requested by the plaintiffs [2, 3].
• Government Intervention: Further filings from the Department of Justice regarding the national security implications of restricting AI training data [3].
• Settlement Activity: Potential movement toward licensing agreements, similar to the $1.5 billion settlement reached between Anthropic and a coalition of authors and publishers last year [2].

The outcome of this case will likely determine whether news organizations can successfully force AI developers to pay for the data used to power their systems or if the "fair use" doctrine will continue to shield the industry's rapid expansion.