Phoenix 24
The legal battle over AI training is moving directly into the newsroom.
New York, United States
USA Today Co. and 13 affiliated entities representing 19 publications have filed a federal lawsuit against OpenAI, accusing the company of using hundreds of thousands of copyrighted articles without permission to train artificial intelligence models. The plaintiffs are seeking more than $250 million in damages and are asking a court to determine whether the alleged infringement was deliberate.
The complaint argues that content from USA Today and several regional newspapers was incorporated into datasets used to train earlier OpenAI systems. The publishers contend that their journalism became part of the technological foundation of commercial AI products without licensing agreements or compensation. They also claim that large-scale copying extended beyond individual stories and reached deeply into archived news content.
A central argument in the case concerns substitution. The publishers allege that AI-generated responses can reduce the need for users to visit original news sites, weakening advertising, subscriptions and direct relationships with readers. Internal statements attributed in the lawsuit to OpenAI employees are being used to support the claim that the company understood AI products could increasingly replace traditional visits to publisher platforms.
The lawsuit also raises a deeper copyright question: whether training an AI model on protected journalism qualifies as fair use. That issue has become one of the defining legal disputes surrounding generative artificial intelligence. Technology companies generally argue that training systems learn patterns from large quantities of information rather than simply reproducing individual works. Publishers and other rights holders counter that commercial AI systems derive economic value from copyrighted material without necessarily compensating those who created it.
USA Today’s action joins a growing wave of litigation from major and regional publishers. The dispute is increasingly dividing the media industry into two approaches. Some organizations are pursuing legal action, while others are negotiating licensing agreements that allow AI companies to use journalistic content under defined commercial terms.
The outcome could influence far more than one company or one newspaper group. Courts may ultimately help determine how intellectual property, journalism and artificial intelligence coexist when models depend on vast quantities of human-created information. The decisions could reshape licensing markets, training practices and the economics of digital publishing.
The conflict therefore reaches beyond copyright. It asks who captures the economic value created when journalism becomes part of an artificial intelligence system, and whether the institutions that produce original reporting can remain sustainable when machines increasingly mediate access to information.
Artificial intelligence may transform how knowledge is delivered, but the battle over who owns its foundations has only begun.