The 19 publications owned by USA Today Co. joined a number of news outlets Thursday when they sued artificial intelligence giant OpenAI for using their copyrighted content without permission to train its programs.
The Louisville Courier Journal is a named plaintiff in the suit, which was filed in New York federal court.
The lawsuit alleges OpenAI’s large language models were trained on the plaintiffs’ copyrighted material "without regard for the rights of publishers and authors.” USA Today Co., formerly known as Gannett, is seeking $250 million in damages from OpenAI for scraping hundreds of thousands of articles from their publications.
“OpenAI’s commercial success rests on large-scale copyright infringement,” the lawsuit reads.
OpenAI has been sued for copyright infringement by several other news organizations in recent years, including the New York Times, whose lawsuit added Microsoft as a defendant this summer. Others who have sued the company include eight newspapers owned by Alden Global Capital Group, such as the Chicago Tribune, Denver Post and New York Daily News.
OpenAI did not immediately respond to a request for comment.
The lawsuit says that USA Today’s publications showed up in over 122 million tokens as of a 2019 snapshot of the Common Crawl, often called “C4,” which is an open internet repository heavily weighted in GPT-3. The Courier Journal’s website originated 3.1 million tokens alone. An exhibit in the lawsuit also points to a specific ChatGPT response in which it mirrors and reproduces the content of an original Courier Journal article.
The USA Today complaint cites recently unredacted documents in the New York Times lawsuit, including a Microsoft executive describing OpenAI’s unauthorized use of copyrighted content as “an astonishing theft of unprecedented proportions” and perhaps the “largest theft of labor in human history.”
The USA Today lawsuit also quotes senior OpenAI executives describing their products as a substitute for the news it copies. ChatGPT leader Nick Turley wrote that news publishers face an “existential threat” from OpenAI’s products, which “will get more and more substitutive as they get better.”
The lawsuit asks for more than money and a jury trial; they request the judge to order the destruction of all large language models and training sets that incorporate USA Today content, and permanently bar OpenAI from using their articles in the future.
*Editor's note: The OpenAI Foundation provides financial support to Louisville Public Media. In accordance with our ethics policy, funding has no influence on our editorial decisions. We do not use generative AI in our reporting.
