A coalition of local and regional newspaper publishers representing nearly 400 U.S. newspapers filed a federal copyright lawsuit against OpenAI and Microsoft on June 24, 2026, alleging the companies illegally used millions of copyrighted articles to train their generative AI models, including ChatGPT and Microsoft Copilot. The complaint, lodged in the Southern District of New York, marks the largest collective action by American news publishers against AI developers to date and intensifies the legal battle over how AI systems ingest and repurpose journalistic content without compensation.

The plaintiffs include a diverse array of community news organizations, from small-town weeklies to major metro dailies, coordinated through a newly formed entity called the Local Newspaper Copyright Alliance (LNCA). They accuse OpenAI and Microsoft of scraping text from their websites, bypassing paywalls, and incorporating verbatim or near-verbatim excerpts into AI-generated responses—all without permission or payment. The lawsuit seeks statutory damages that could exceed billions of dollars if the court finds willful infringement, as well as an injunction halting further use of their content.

The Allegations: Systematic Scraping and Unfair Competition

At the heart of the lawsuit is the claim that OpenAI and Microsoft built their lucrative AI businesses on a foundation of stolen intellectual property. The newspapers argue that the companies copied massive datasets—collectively known as a training corpus—by crawling the open web and hoovering up articles en masse. The complaint highlights that many of the publishers employ paywalls or subscription models, yet their content was nonetheless ingested. Some articles allegedly appear nearly verbatim in ChatGPT and Copilot outputs, effectively becoming a substitute for visiting the original source.

“This is not a case about technology versus tradition,” the complaint states. “It is about two of the world’s richest companies misappropriating the lifeblood of local journalism—original reporting—to build a competing information service that drains readership and revenue from the very newsrooms that produce that reporting.” The publishers also claim that Microsoft’s Copilot, integrated directly into Windows 11 and Microsoft 365, exacerbates the harm by placing AI-summarized news directly within the operating system, further diverting traffic from their websites.

Microsoft’s Deep Integration of Copilot in Windows

For Windows users, the lawsuit shines a harsh light on one of Microsoft’s most ambitious product integrations. Since the release of Windows 11 24H2, Copilot has been deeply woven into the desktop experience—appearing as a sidebar, a taskbar icon, and even within native applications like Notepad and Paint. When users ask Copilot for news summaries or background on current events, the AI draws upon its training data, which the plaintiffs argue includes their copyrighted material. This means that a Windows user might receive a concise breakdown of a local city council meeting, sourced from a newspaper’s original reporting, without ever seeing the publisher’s website or advertisements.

Microsoft has positioned Copilot as a productivity tool that “empowers every person on the planet to achieve more.” But for newspaper executives, that empowerment comes at their expense. “Every time Copilot answers a question with our content, it takes a potential subscriber away from us,” said one publisher involved in the suit, who requested anonymity to discuss legal strategy. “Microsoft is essentially building a replacement for local news inside Windows, and we get nothing in return.”

The Windows connection also raises technical questions about how Copilot retrieves information. While Microsoft has emphasized that Copilot in Windows sometimes uses Retrieval-Augmented Generation (RAG) to fetch real-time data from the web, the model’s base training still relies on the pre-existing corpus. The lawsuit targets both the initial training and any ongoing scraping or caching that might occur.

The LNCA filing is the latest—and largest—in a growing docket of lawsuits pitting content creators against AI giants. In late 2023, The New York Times sued OpenAI and Microsoft, setting off a cascade of similar actions. Book authors, visual artists, and music publishers have all filed claims. But the local newspaper coalition represents a different scale: unlike a single national publication, these hundreds of outlets collectively form the backbone of civic information in communities across America. If they prevail, the financial and operational impact on AI development could be seismic.

Previous cases have hinged on the fair use doctrine, which allows limited use of copyrighted material without permission for purposes such as criticism, comment, news reporting, teaching, or research. OpenAI and Microsoft have consistently argued that training AI models constitutes transformative fair use because the system does not replicate works wholesale but rather learns patterns to generate new, original outputs. However, the newspapers point to numerous instances where ChatGPT or Copilot reproduced lengthy passages from their articles, sometimes with only minor paraphrasing. They contend this is not transformative but plain reproduction.

Fair Use Under the Microscope

Legal scholars are divided. Some believe courts will tilt toward the AI companies, citing precedent that favors innovative technologies and the sheer impracticality of licensing millions of individual works for training. Others say the scale and commercial nature of the AI products, combined with direct evidence of memorization and regurgitation, could overcome the fair use defense. The U.S. Copyright Office has also signaled that it plans to address AI and copyright in an upcoming report, which could influence judges.

One key factor is the effect on the potential market for the original work. If a Copilot news summary replaces the need to visit a newspaper’s site, that directly undermines the paper’s subscription and advertising model. The plaintiffs will argue that OpenAI and Microsoft are not merely using the content for research but are creating a competing product. The defendants will counter that AI outputs are not substitutes because they lack the depth, editorial oversight, and reliability of original journalism.

The Fate of Local Journalism Hangs in the Balance

Local newspapers have been in sharp decline for two decades, battered by falling print circulation and the rise of digital platforms. According to the Medill Local News Initiative, the U.S. has lost over 2,500 newspapers since 2005, creating “news deserts” where residents lack reliable information about their communities. The publishers behind this lawsuit argue that AI scraping is accelerating that trend by siphoning what little digital revenue remains.

“We can’t compete if our own work is used against us,” said a statement from the LNCA. “Every journalist, every editor, every photographer—their labor is being exploited to enrich Microsoft and OpenAI. This isn’t innovation; it’s predation.” The alliance has set up a legal fund and is calling on Congress to clarify that using copyrighted material for commercial AI training requires a license.

For Windows users, the dispute could eventually reshape how Copilot operates. If the court grants an injunction, Microsoft might be forced to remove training data derived from the plaintiffs’ content and implement filtering mechanisms to prevent its future use. That could make Copilot less knowledgeable about local affairs or require a redesign to rely solely on licensed data sources. In the short term, however, the lawsuit is likely to drag on for years, with appeals stretching any resolution into the next decade.

Microsoft and OpenAI’s Response

Neither company has formally responded to the LNCA complaint, but both have vigorously defended themselves in similar cases. Microsoft has argued that training on public web data is a long-standing industry practice, akin to how search engines index web pages. OpenAI has struck licensing deals with some publishers—including the Associated Press and Axel Springer—while maintaining that it is not legally required to do so for all training data. A spokesperson for OpenAI declined to comment on pending litigation but pointed to those agreements as evidence of the company’s willingness to support journalism.

Privately, some tech executives view the lawsuits as an existential threat to generative AI. If every website owner could demand royalties for inclusion in training sets, the cost of producing models like GPT-6 or Copilot Next could become prohibitive. This reality has spurred behind-the-scenes discussions in Washington about a possible statutory license or compulsory royalty system, similar to what exists in the music industry.

What This Means for Windows and Copilot Enthusiasts

For the millions of people who use Windows daily, Copilot’s future hangs in a delicate balance. Microsoft has touted Copilot as the future of computing, weaving it into search, email, word processing, and even security. A court ruling that forces it to sever ties with vast swaths of the internet’s text could dull the assistant’s utility. Windows Insiders in the Dev Channel may start seeing experimental builds that test new approaches to content attribution or filtering, but no official changes have been announced.

In the meantime, users can expect Copilot to continue functioning as it does today, but with a growing legal asterisk. If you ask Copilot about a local event and it summarizes a newspaper’s reporting, that summary may one day be attributed to a licensed source—or disappear entirely. The outcome of this case will shape not just the fate of local news but the fundamental architecture of AI assistance across the Windows ecosystem.

The LNCA lawsuit is on track to become a landmark case at the intersection of copyright, technology, and press freedom. Its sheer scope—nearly 400 plaintiffs from all corners of the country—gives it political weight that earlier, single-plaintiff suits lacked. Congress may be forced to act if the courts cannot settle the matter. For now, the judge assigned to the case, who has yet to issue any preliminary rulings, will sift through thousands of pages of legal briefs in the coming months.

Windows enthusiasts watching this space should pay close attention to any settlement talks. A negotiated resolution could involve Microsoft and OpenAI paying into a news fund, much like Google’s News Showcase, while maintaining broad access to content. Alternatively, a sweeping court victory for the newspapers could force the AI industry to rethink its entire data acquisition strategy.

Whatever happens, one thing is clear: the long-running tension between technology platforms and the news media has entered a new, more acrimonious chapter. And this time, the battleground is code rather than advertising dollars.