USA TODAY Co. and affiliated newspaper companies sued OpenAI in the US District Court for the Southern District of New York on 8 October 2026, alleging that OpenAI copied their journalism to train its GPT models without permission and seeking damages in excess of $250 million.
The complaint says the plaintiffs own copyrights in content published by 19 publications, including USA TODAY, The Tennessean, the Detroit Free Press and The Arizona Republic. It brings three counts: copyright infringement, vicarious copyright infringement, and removal of copyright management information under the Digital Millennium Copyright Act.
Business Pill · A CLAIM IS NOT A DECISION
A short explainer of allegations: a claim made in a filing is not a decision by a court. It teaches the general idea only and says nothing about any party in this story.
The key insight: As we read it, the case rests mainly on alleged copying, from dataset counts to memorization and removed copyright information, and it also points to OpenAI’s licensing deals as evidence that OpenAI knows a license is required. Its remedies go beyond damages to destruction of models trained on the plaintiffs’ content.
The plaintiffs allege that their content comprises more than 160,000 entries in OpenAI’s WebText dataset, including 83,266 entries from usatoday.com, and over 122 million tokens in C4, a filtered subset of a 2019 Common Crawl snapshot.
They allege that OpenAI’s models “memorized” the works they ingested, and that OpenAI used programs designed to strip copyright management information from them. The complaint says the law allows up to $150,000 for each willful infringement plus up to $25,000 per violation for removing that information.
The complaint also cites responses from GPT-5.6 that, it says, retrieved articles from its newspapers and produced extensive summaries following the originals’ structure; full examples are in an exhibit.
Entries in OpenAI’s WebText dataset per plaintiff domain, as alleged in paragraph 102 of the complaint filed on 8 October 2026: usatoday.com 83,266, freep.com 12,994, jsonline.com 9,814, azcentral.com 9,073 and others. These are allegations, not findings.What the plaintiffs seek, per the complaint filed on 8 October 2026: damages in excess of $250 million, an injunction and destruction of models and training sets. These are requests, not findings.
Licensing and the Wider Litigation
The plaintiffs argue that OpenAI knows a license is required, pointing to licensing agreements it says OpenAI has with over a dozen news and media organizations, including the Associated Press, Axel Springer, The Atlantic and Vox Media.
The complaint says the plaintiffs join a long list of copyright holders suing OpenAI and other AI companies, many of whose cases have been consolidated in the same court. It names models from GPT-1 through GPT-6.1 and GPT-OSS among those at issue.
Beyond damages, the plaintiffs ask for a permanent injunction and for destruction of GPT or other models and training sets that incorporate their content, and they demand a jury trial.
The Structural Read
The evidence is set out at dataset level. The plaintiffs count their domains’ entries in WebText and tokens in C4, and the complaint says OpenAI has kept the training data of its later commercial models secret.
Retrieval is part of the case. The complaint cites GPT-5.6 summaries of its newspapers’ articles and argues that such results give readers less reason to visit the original source.
It joins an existing docket. The complaint says many similar cases against OpenAI and other AI companies have been consolidated in the same court.
USA TODAY Co. et al. v. OpenAI, complaint, 8 October 2026
“Those agreements prove that OpenAI understands copyrighted journalism has value and that using it requires permission and payment.”
Three Implications
LICENSING AS EVIDENCE The complaint cites OpenAI’s agreements with the Associated Press, Axel Springer, The Atlantic and Vox Media.
MODELS, NOT ONLY MONEY The plaintiffs ask for destruction of GPT or other models and training sets that incorporate their content.
RETRIEVAL IN SCOPE The complaint cites GPT-5.6 summaries of articles from its newspapers as examples.
The framework’s starting point: “Close the SCARCE layer (keep proprietary) + Open the ABUNDANT layer (commoditize).”
As we read it, the suit is a publisher arguing that its archive belongs in the scarce, licensed layer rather than among freely available inputs; the complaint points to OpenAI’s licensing agreements with other news organizations as evidence that such content has a price. Whether a court agrees is not established.
What Is Not Established
These are allegations in a complaint filed on 8 October 2026, and no court has ruled on them. When we checked the court docket, its entries were the plaintiffs’ filings and court notices, with no response from OpenAI. We read the complaint as filed but not its exhibits. We did not contact the parties.
USA TODAY Co. and its affiliates allege that OpenAI trained on and reproduces journalism from 19 of their publications, citing more than 160,000 WebText entries, and seek over $250 million, an injunction and destruction of models trained on their content.
A note on sourcing. We read the complaint filed on 8 October 2026 in the Southern District of New York (No. 1:26-cv-08892) and the court docket; we did not read the exhibits. Every figure is the plaintiffs’ allegation. We did not contact the parties. Nothing here is a forecast, and nothing here is financial or investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
