A publisher coalition suing OpenAI and Microsoft reportedly grew past 550 outlets on September 16, 2026, after 26 publishers joined the group. Platkin LLP reportedly represents 60 publishers directly. The development expands the number of publishers pressing allegations over the use of copyrighted journalism in AI training, model development and outputs, but it does not decide whether OpenAI or Microsoft infringed copyright.
The publisher coalition reportedly passes 550 outlets
The 26 new members reportedly included Times Publishing Company, Austin Chronicle Corp., Alternative Newsweekly Foundation and SwimSwam Partners. The coalition says OpenAI and Microsoft used copyrighted news articles without permission or payment to build commercial AI products. It also alleges that reporting was reproduced or repurposed without compensating the news organizations that produced it.
Another allegation concerns copyright-management information. The Platkin-related plaintiffs say OpenAI removed author credits, copyright notices and terms-of-use data, and characterize that conduct as a violation of the Digital Millennium Copyright Act.
The coalition's reported size and the claims attached to it remain part of active litigation. They are not a finding of liability.
What publishers allege against OpenAI and Microsoft
The disputes cover more than one use of copyrighted material. Publishers and authors have raised claims involving the copying of works for model training, the production of outputs that may reproduce protected expression, contributory infringement and the handling of information identifying copyright owners.
The New York Times Company filed a copyright lawsuit against Microsoft and OpenAI on December 27, 2023. It alleged that millions of Times articles were used to train AI systems and argued that those systems could compete with the newspaper as an information source.
On April 30, 2024, eight U.S. newspaper publishers filed another action alleging unauthorized use of millions of articles in ChatGPT, Copilot and related systems. OpenAI and Microsoft have denied the consolidated complaint's allegations and argued that using publicly available data can qualify as fair use and transform the underlying works into new material.
That defense matters because fair use is not a blanket answer for every use of a copyrighted work. The court must consider the circumstances of the use, including how the material was used, the nature of the copyrighted work, the amount involved and the effect on the market for the original.
The coalition is separate from the 12 consolidated cases
The reported 550-plus outlets and the 12 cases consolidated in the Southern District of New York in April 2025 describe different things. The first is the reported size of a publisher coalition. The second is the number of related copyright cases brought by authors and news organizations that were combined into a multidistrict proceeding.
The consolidated litigation is overseen by District Judge Sidney H. Stein and Magistrate Judge Ona T. Wang. The case history includes procedural decisions, but those decisions do not amount to a final ruling that OpenAI or Microsoft infringed copyright.
For example, on October 27, 2025, the court denied OpenAI's motion to dismiss an output-based direct-infringement claim in the Alter litigation. The ruling allowed that claim to continue past the dismissal stage without deciding the fair-use question.
A December 15, 2025 ruling described in the case record allowed certain Digital Millennium Copyright Act section 1202(b) and contributory-infringement claims in the Ziff Davis litigation to proceed, while dismissing the section 1202(a) claim described in that record. These are pleading-stage outcomes, not a final determination of infringement.
The central question is what fair use covers
At the heart of the cases is whether copying copyrighted works to train AI models, and generating outputs that reproduce protected expression in some circumstances, is legally protected fair use. The answer could differ across claims and factual settings rather than turning on the number of publishers involved.
A September 4, 2026 report said the parties had filed competing summary-judgment motions concerning fair use. The proceedings therefore have moved beyond the first wave of complaints and dismissal motions, while the legal question remains contested.
Fred Havemeyer, head of U.S. AI and software research at Macquarie, has argued that copyright law will help define how the generative-AI ecosystem evolves. That pressure reaches beyond the courtroom: publishers are seeking compensation or licensing arrangements, while AI companies are defending the use of publicly available data and the transformative nature of model training.
What changes for AI companies and publishers
For publishers, the litigation tests whether valuable reporting can be used to build commercial AI systems without a licensing arrangement or payment. The claims also challenge how much of an article's expression can appear in an output and whether author or copyright information survives the process.
For OpenAI and Microsoft, the cases put training practices, model outputs and the fair-use defense under the same broad legal spotlight, even though those issues involve separate claims. A decision on one procedural motion would not automatically resolve every allegation in the wider litigation.
The next important legal outcomes will come from the courts' handling of the fair-use proceedings and the individual claims that have survived dismissal. The reported coalition expansion changes the scale of the plaintiffs' campaign; it does not by itself decide the copyright question.
