Two regional US newspapers are asking a federal judge to order the destruction of AI training sets and models built from their reporting. The Seattle Times Co. and Newsday lodged the copyright and trademark complaint in the Southern District of New York on September 4, a filing their own newsrooms said runs 38 pages.
OpenAI and Microsoft are the defendants. The papers claim their journalism was ingested to train and operate AI products without permission or payment, and that the models reproduce it, often verbatim. Fabricated content falsely attributed to the two brands amounts to trademark dilution, the outlets add. Court papers cite industry numbers showing search referrals to midsize publishers down 47% year over year last December as evidence of the harm.
Beyond unspecified damages, the publishers want the impoundment or destruction of datasets and models containing their articles. Seattle Times chief executive Alan Fisco told staff in an email that the call was not easy, and that the company must defend material it spends millions of dollars a year producing. The case, he said, is not meant to block AI progress but to keep it from riding on the newspaper’s business model.
