Boone Newsmedia, Inc., an Alabama company that publishes community newspapers in seven states, has sued Microsoft and 10 OpenAI entities for copyright infringement in federal court in Mississippi. The complaint, filed Oct. 30, 2026, accuses the companies of copying “tens of thousands of copyrighted articles” to build ChatGPT and Microsoft Copilot without permission or payment.
The filing centers on how those articles allegedly ended up inside AI systems, and on what Boone says it can and cannot show. Here are the basics from the complaint and the court docket.
| Item | Detail |
|---|---|
| Case | Boone Newsmedia, Inc. v. Microsoft Corporation, No. 3:26-cv-00950 |
| Court | U.S. District Court for the Southern District of Mississippi, Northern Division |
| Filed | Oct. 30, 2026, with a jury demand |
| Defendants | Microsoft Corporation; OpenAI, Inc.; OpenAI LP; OpenAI GP, LLC; OpenAI, LLC; OpenAI OpCo, LLC; OpenAI Global, LLC; OAI Corporation, LLC; OpenAI Holdings, LLC; OpenAI Foundation; OpenAI Group PBC |
| Claims | Direct copyright infringement, vicarious copyright infringement, and removal of copyright management information under the Digital Millennium Copyright Act |
| Plaintiff’s lawyers | Barrett Law Group; Wilson Carroll, PLLC; Cuneo Gilbert Flannery & LaDuca, LLP |
What follows is Boone’s account. None of the claims has been proven in court.
The publisher behind the suit
The complaint describes Boone as an Alabama corporation headquartered in Selma that “owns and/or manages dozens of newspapers,” along with websites, shopping guides and magazines. It names communities in Alabama, Georgia, Michigan, Minnesota, Mississippi, Virginia and North Carolina.
The company’s own history page puts the count at 29 newspapers. It traces Boone’s origins to Tuscaloosa Newspapers, Inc., which leased The Tuscaloosa News from 1951 until 1981, and says founder James B. “Jim” Boone Jr. died in February 2023.
The newspapers named in the filing
The complaint lists the Boone companies and their websites by web address. By our count, it names 31 websites under 18 company names. Shelby County Newspapers Inc. accounts for nine of them, including the Shelby County Reporter and magazine sites named for Hoover, Mountain Brook, Vestavia Hills, Homewood, Leeds, Moody and Trussville.
Newspaper sites on the list include the following, with names and places matched from the web addresses and Boone’s website directory:
- Alabama: The Clanton Advertiser, The Greenville Advocate, The Lowndes Signal, The Luverne Journal, The Selma Times-Journal, the Troy Messenger and The Valley Times-News
- Georgia: LaGrange Daily News
- Michigan: The Argus-Press
- Minnesota: Albert Lea Tribune and Austin Daily Herald
- Mississippi: The Natchez Democrat and The Vicksburg Post
- North Carolina: The Coastland Times, the Roanoke-Chowan News-Herald, the Gates County Index and the Washington Daily News
- Virginia: The Smithfield Times, the Suffolk News-Herald, The Tidewater News and Windsor Weekly
How the complaint says the articles were taken
According to the complaint, automated systems crawled Boone’s websites, “including content behind paywalls,” and copied the articles onto the defendants’ servers. It says the copies fed training datasets such as Common Crawl, WebText and WebText2, which OpenAI used for its GPT models.
The filing offers three data points. It says an analysis of Common Crawl shows that tens of thousands of Boone articles, and tens of millions of tokens of its content, likely appeared in the version used to train GPT-3.
A consultant for Boone’s lawyers, it says, found thousands of records from Boone sites in OpenWebText, an open-source stand-in for WebText. The complaint also says the C4 dataset holds full Boone articles stripped of bylines, titles and copyright notices.
That stripping is the basis of the Digital Millennium Copyright Act count. The complaint alleges OpenAI used two text-extraction tools, Dragnet and Newspaper, that separate article text from author names, copyright notices and terms of use. It alleges, on information and belief, that OpenAI picked the tools for those qualities.
The complaint casts Microsoft as OpenAI’s partner. It says Microsoft invested about $13 billion in OpenAI, built a dedicated Azure supercomputer for OpenAI’s training and serves the models through Copilot.
On willfulness, the complaint quotes OpenAI’s 2023 written evidence to a U.K. House of Lords committee, which said it would be “impossible to train today’s leading AI models without using copyrighted materials.”
No Boone output example in the filing
According to the complaint, earlier publisher suits against OpenAI were backed by side-by-side comparisons of chatbot answers and original articles. The Boone complaint does not include one for its own work. Our review of all 37 pages found no ChatGPT or Copilot response quoting a specific Boone story.
Instead, the filing leans on testing by publishers that sued first. It names The New York Times, the Chicago Tribune, The Denver Post, The Intercept, Ziff Davis and others, and says their tests showed GPT models producing “hundreds of words of verbatim text” from single articles.
The complaint then alleges, “upon information and belief,” that OpenAI has since disabled public access to the ChatGPT versions those publishers tested. It argues those versions “would likewise generate near-verbatim copies” of Boone’s articles. Further details about output of Boone content, it says, are “solely in OpenAI’s possession, custody, and control.”
The complaint says Boone “is the owner of registered copyrights,” but it does not list registration numbers or specific articles. The docket shows only a civil cover sheet attached to it.
What Boone is asking the court to order
The complaint does not name a dollar amount. It asks for statutory damages, including for willful infringement, along with compensatory damages, restitution and disgorgement of profits.
It also seeks a permanent injunction against the conduct described in the complaint. Under a separate section of the Copyright Act, it asks for an order requiring the defendants to “remove all copies of Registered Works” from all GPT or other large language models and their training sets. Costs, attorney’s fees and a jury trial round out the requests.
A second Mississippi case and the docket so far
A CourtListener search of the same court’s records turned up a similar case. Emmerich Newspapers, Inc. v. Microsoft Corporation, No. 3:26-cv-00791, was filed there on Oct. 2, 2026, against the same 11 defendants.
That suit was filed by Emmerich Newspapers and five Coopwood media companies. Boone’s complaint does not mention the Emmerich case.
As of Nov. 3, CourtListener’s copy of the Boone docket showed three entries, all dated Oct. 30: the complaint, summonses to all 11 defendants and a notice of the suit mailed to the U.S. Copyright Office. The docket showed no response from Microsoft or any OpenAI entity and no ruling from the court. The Boone Newsmedia lawsuit is at its earliest stage.
Featured image: The Selma Times-Journal building in Selma, Alabama. The paper is one of the Boone sites named in the complaint. Photo: Michael Barera / Wikimedia Commons (CC BY-SA 4.0)
