The complaint, filed on Thursday in the US District Court for the Southern District of New York, accuses OpenAI of copying hundreds of thousands of copyrighted articles without permission or payment. The publishers are also seeking a permanent injunction against the alleged infringement and other remedies involving the company’s artificial intelligence models.
The action names seven OpenAI entities as defendants and involves USA Today alongside 18 affiliated publications. The plaintiffs allege that OpenAI systematically collected their reporting, incorporated it into training datasets and subsequently developed commercial products capable of reproducing protected material.
The publishers contend that the alleged copying deprived them of licensing revenue while enabling OpenAI to build products that compete with established news organisations for readers and digital traffic. Those assertions remain allegations before the court, and no determination of infringement has been made.
The lawsuit also alleges that OpenAI reproduced or repackaged copyrighted journalism in responses generated by its artificial intelligence systems. According to the complaint, some outputs omitted information identifying the original publishers and copyright holders, raising separate questions under federal copyright legislation.
The plaintiffs have requested damages exceeding $250 million, recovery of profits attributable to the alleged infringement and an order preventing further unauthorised exploitation of their material. They additionally seek destruction of models and training datasets containing their copyrighted works, an unusually extensive remedy whose availability would require judicial determination.
OpenAI had not publicly responded to the specific allegations when the lawsuit was filed. The company has maintained in other copyright disputes that training artificial intelligence systems on publicly accessible material can constitute fair use under US law, particularly when the resulting technology transforms information into new applications.
The publishers reject that position as applied to their journalism. Their complaint argues that copying protected expression on a substantial scale for commercial model development cannot automatically qualify for copyright protection merely because the resulting technology performs different functions.
A central allegation concerns the datasets used to develop OpenAI’s language models. The complaint identifies more than 160,000 entries associated with publisher domains in the WebText corpus, which the plaintiffs cite as evidence of extensive collection of their material.
The lawsuit further alleges that OpenAI’s systems can generate passages resembling protected articles, potentially allowing users to obtain journalistic content without visiting the original publications. Whether particular outputs reproduce legally protected expression, rather than unprotected facts, will be among the questions requiring examination.
The plaintiffs include publishers associated with The Tennessean, Detroit Free Press, The Arizona Republic, The Columbus Dispatch, The Des Moines Register, Milwaukee Journal Sentinel, The Oklahoman and The Courier-Journal. Their participation brings national and regional newspaper operations into the same legal challenge.
The complaint sets out claims involving direct copyright infringement, vicarious infringement and the alleged intentional removal of copyright management information under the Digital Millennium Copyright Act. Each claim involves distinct legal requirements, including questions about copying, responsibility and the handling of identifying information.
The case adds to litigation confronting artificial intelligence developers over the material used to train large language models. Publishers and authors have argued that commercial exploitation of copyrighted works requires authorisation, while technology companies have invoked fair use and the transformative characteristics of machine learning.
OpenAI and Microsoft, its principal financial backer, already face copyright claims brought by The New York Times and other rights holders in Manhattan federal court. Those proceedings address overlapping questions concerning training datasets, generated outputs and the economic implications of artificial intelligence systems.
The new complaint seeks to establish liability for the use of journalism across multiple stages of artificial intelligence development, from acquiring source material to producing responses for users. It also raises the question of whether alleged infringement extends beyond initial training into the operation of deployed products.
The publishers have requested a jury trial. Their legal representatives include Steven Lieberman of Rothwell, Figg, Ernst & Manbeck, while the proceedings are registered as USA Today Co. Inc. versus OpenAI Foundation, case number 1:26-cv-08892.
Court records identify the filing date as October 8, 2026. The complaint was accompanied by supporting exhibits addressing copyright registrations and examples of generated outputs, materials that the plaintiffs intend to use in substantiating their allegations before the federal court.
Follow Arabian Post
Select Arabian Post as your preferred source on Google and MSN News for trusted business news and Arab politics and updates.
