USA Today's Owner Sues OpenAI Over Its Journalism

The company behind USA Today has taken OpenAI to court, accusing it of training its models on hundreds of thousands of copyrighted articles without permission and then building a product that discourages readers from ever visiting the source. The suit, filed on October 8 in the Southern District of New York, seeks more than $250 million in damages and an order to halt the alleged infringement.

The plaintiff is USA Today Co., the publisher formerly known as Gannett and the largest newspaper chain in the United States. Its complaint, reported by Unite.ai and Crypto Briefing among others, covers work from 19 titles, including USA Today, The Arizona Republic, the Detroit Free Press, and the Indianapolis Star. It names the OpenAI Foundation and several affiliated entities as defendants.

What gives the filing its edge is the second half of the argument. USA Today Co. does not only object to how its journalism was used in training. It says ChatGPT's answers now stand in for the articles themselves, summarizing reporting closely enough that readers have little reason to click through. The complaint quotes OpenAI's own head of ChatGPT conceding that once the chatbot has given an answer, there is "no good reason to click" a link to the underlying source. For a business that lives on readers arriving at its pages, that is the whole fight in a sentence.

Another front in a widening war

This is a complaint, not a verdict, and its claims have not been tested in court. OpenAI had not responded publicly when the first reports ran. The company has argued in other cases that training on published material can qualify as fair use, a position USA Today Co. squarely rejects for its own content.

The case joins a crowded docket. The New York Times, The Intercept, Ziff Davis, and a long list of other publishers have already sued OpenAI or its partners, and the outcomes so far have cut both ways. Late last month we reported that a federal appeals court rejected a fair-use defense in a related dispute, a ruling that will have given newspaper lawyers some encouragement. At the same time, OpenAI's business keeps growing, with revenue approaching a $70 billion annual run rate, which means it has both the motive to keep scraping and the means to litigate for years.

The deeper question is one courts have been circling for two years without settling. If a model reads the news, learns from it, and then answers questions that would once have sent a reader to a newspaper, who owns the value that changes hands? USA Today Co. has put a number on its answer. Now a judge has to decide whether the law agrees.