The legal noose is tightening around OpenAI. The Seattle Times and Newsday have filed separate lawsuits accusing the AI giant of copyright infringement, alleging it scraped their journalism to train models without consent. This isn’t an isolated incident. It mirrors high-profile cases brought by The New York Times, Ziff Davis, and Merriam-Webster. The core complaint remains consistent: unauthorized use of proprietary content that now competes directly with the original sources.
Microsoft is dragged into the fray as a co-defendant. Since its Copilot service relies on OpenAI’s underlying technology, the publishers argue the software giant shares liability. The stakes extend beyond mere attribution. These outlets claim chatbots provide answers that eliminate the need for users to visit their websites. That direct substitution erodes subscription revenue, striking at the financial heart of modern journalism. It is a survival issue, not just a principle.
The remedies sought are drastic. Both plaintiffs demand the destruction of any copies of their works held by the defendants. They also want the associated training datasets and AI models purged of their content. This goes beyond seeking damages. It aims to dismantle the infrastructure built on their intellectual property. Such a ruling could force a massive retraining effort or limit model capabilities significantly. OpenAI and Microsoft have not yet commented on these specific filings.
This litigation joins a broader wave involving nearly 400 local newspapers. The sheer volume of plaintiffs suggests a coordinated industry pushback rather than scattered grievances. While tech firms often cite fair use, courts may view the commercial replication of entire articles differently. If judges side with the publishers, the cost of building large language models could skyrocket. We might see a future where AI companies must negotiate licenses for every byte of training data, fundamentally changing how these systems are built.
