Particle.news

Seattle Times and Newsday Sue OpenAI and Microsoft Over Alleged Use of Paywalled Reporting

A court order forcing destruction of copies or models could redefine how AI firms collect and use news for training purposes.

Overview

  • The Seattle Times and Newsday filed federal suits in the Southern District of New York that accuse OpenAI and Microsoft of scraping their websites, including paywalled articles, and using that content to train products such as ChatGPT, Copilot and Bing’s AI.
  • The newspapers asked the court to require destruction of unauthorized copies, the training datasets that contain their work, and any models that incorporate the material.
  • OpenAI and Microsoft deny wrongdoing, say their models use publicly available data and that training can be protected by fair use, and both said they are open to talks with the publishers.
  • The filings follow related litigation, including a 2023 New York Times case in which the U.S. Department of Justice recently argued that training on copyrighted material can be fair use, a position that could influence judges weighing these new requests.
  • If a judge orders deletion of data or models, the ruling could force changes to how companies gather web content, prompt new licensing deals for publishers, and directly affect local newsrooms that say AI outputs have reduced site visits and subscription revenue.