USA Today sues OpenAI over copyright, seeks more than $250M

USA Today Co. and 13 affiliated newspaper entities sued OpenAI on Oct. 8 in the Southern District of New York, alleging the company trained GPT models on hundreds of thousands of their articles without permission, per Forbes and Unite.AI. The complaint says the papers’ content appears more than 160,000 times in WebText, the dataset behind GPT-2, and that a 2019 Common Crawl snapshot used to train later models pulled 23 million tokens from usatoday.com alone. Beyond copyright and vicarious-infringement claims, the suit adds a DMCA count over stripped authorship credits, seeks more than $250 million including up to $150,000 per willfully infringed work, and asks the court to fold the case into OpenAI’s existing consolidated copyright litigation.