USA Today Co. and more than a dozen of its newspapers sue OpenAI, seeking $250 million over articles used to train ChatGPT

USA Today Co., the newspaper group formerly known as Gannett, sued OpenAI on Thursday 8 October, Hoodline and TheDesk.net reported. TheDesk.net says more than a dozen of the company's papers are plaintiffs, among them The Arizona Republic, the Milwaukee Journal Sentinel, the Des Moines Register and The Tennessean. Hoodline says the case was filed in federal court in Manhattan.

The publishers accuse OpenAI of copying their articles without permission to train the models behind ChatGPT. Damages sought are more than $250 million, Hoodline reported; TheDesk.net gave the figure as $250 million. Hoodline reported that they also want a court order stopping the alleged infringement and the destruction of any GPT models and training sets built on their work.

Both outlets report the complaint's main evidence: more than 160,000 entries from the group's websites in OpenAI's WebText training dataset. The complaint also argues that ChatGPT's summaries take readers, and so advertising money, away from the papers' own sites, TheDesk.net reported. OpenAI had not commented, the outlets said.

The complaint itself could not be read here, so the figures are carried as reported by two named outlets. The claims are allegations and the court has not ruled. The Switchboard noted the filing in one line on 9 October; this is the full account. It adds USA Today Co. to the publishers already suing OpenAI over training data.

• Reported USA Today Co. and its newspapers sued OpenAI in Manhattan federal court, seeking more than $250 million, Hoodline reported; TheDesk.net gave $250 million. Hoodline
• Reported The suit asks for the destruction of any GPT models and training sets built on the plaintiffs' work, Hoodline reported. Hoodline
• Reported The complaint says OpenAI's WebText dataset holds more than 160,000 entries from the group's sites, TheDesk.net reported. TheDesk.net