USA Today Co. and 13 of the newspapers it owns sued OpenAI in Manhattan federal court, accusing the AI giant of illegally copying hundreds of thousands of articles from 19 publications to build and run ChatGPT. The lawsuit, filed Thursday, seeks more than $250 million in damages, a court order blocking further alleged infringement, and the destruction of any GPT models and training sets built on the plaintiffs' work.
The complaint, reported by Reuters, alleges OpenAI used USA Today and its sister newspapers' content without permission to train the large language models behind ChatGPT. According to the complaint, the company copied hundreds of thousands of articles and other materials, and OpenAI did not immediately respond to a request for comment on the allegations, per that report. The case, USA Today Co. v. OpenAI Foundation, is pending in the U.S. District Court for the Southern District of New York under case number 1:26-cv-08892.
Named Publications and Alleged Token Counts
The suit names work from 19 publications, including USA TODAY, The Tennessean, the Detroit Free Press, The Arizona Republic, The Indianapolis Star, The Columbus Dispatch and The Milwaukee Journal Sentinel, according to runtimewire.com. Articles from the USA Today plaintiffs allegedly make up more than 160,000 entries in a dataset called WebText and over 122 million tokens in a 2019 Common Crawl snapshot known as C4, according to Forbes. Of that total, 23 million tokens reportedly came from usatoday.com alone in that same 2019 snapshot.
The requested damages break down into two categories, per Forbes: up to $150,000 for each work OpenAI allegedly infringed willfully, and up to $25,000 for every instance in which the company is accused of stripping out copyright management information. The plaintiffs also argue that ChatGPT can reproduce or summarize their articles in ways that substitute for readers actually visiting the original publications, the same report from runtimewire.com notes.
Attorneys and the Road Ahead
Steven Lieberman, an attorney at Rothwell Figg Ernst & Manbeck, represents USA Today in the case, per Reuters. Attorney information for OpenAI was not yet available at the time of the Reuters report, and the court had not ruled on any of the plaintiffs' allegations as of the filing, since they remain claims in a newly filed complaint.
Part of a Growing Legal Pile-Up
The USA Today complaint lands amid dozens of high-stakes copyright lawsuits against technology companies, with copyright owners suing OpenAI, Anthropic and Meta Platforms over alleged misuse of their work in AI training, Reuters reported. OpenAI and Microsoft already face similar copyright claims in Manhattan federal court from authors and news outlets, including The New York Times, which sued both companies in December 2023 over claims of copyright infringement, trademark issues, misappropriation and false attribution tied to its reporting.
OpenAI is also facing copyright suits from The Intercept, Ziff Davis, CBC/Radio-Canada, Encyclopaedia Britannica, Merriam-Webster, The Seattle Times and a coalition of nearly 400 local newspapers, according to The Verge. U.S. News & World Report filed its own complaint against multiple OpenAI entities in the same Southern District of New York in late November 2025, and that case was later accepted as related to a broader OpenAI multidistrict litigation proceeding, court records on CourtListener show. Not every publisher has gone the litigation route, however; OpenAI struck a 2024 licensing deal with News Corp covering current and archived content from outlets such as The Wall Street Journal and the New York Post.
How Courts Have Ruled So Far
Technology companies have consistently argued that their systems make fair use of copyrighted material by transforming it into something new, and some court rulings have sided with that defense. In Kadrey v. Meta, a court found AI training transformative and ultimately granted Meta summary judgment on its fair-use defense, according to a 2026 update from Ohio State University's library. In Bartz v. Anthropic, a court similarly found that training AI on legally acquired copyrighted works was fair use, though it denied summary judgment over Anthropic's alleged use of pirated copies to build a central library.
Other rulings have cut the other way. A court granted Thomson Reuters summary judgment after finding that Ross Intelligence's use of Westlaw headnotes to train its AI legal-research tool was not fair use, according to an update from Norton Rose Fulbright. OpenAI has argued more broadly that training models on publicly available internet material is fair use, a position it continues to press as its legal exposure mounts across multiple fronts.
