In a landmark legal challenge, the world’s oldest encyclopedia and dictionary publisher alleges that OpenAI copied nearly 100,000 articles without permission, and is now eating its lunch.
For 258 years, Encyclopedia Britannica has been the gold standard of reliable, curated knowledge, the kind of authoritative reference that students, scholars, and curious minds turn to for verified facts. Now, the venerable publisher is fighting to protect that legacy in a Manhattan federal courtroom, claiming that OpenAI quietly helped itself to the fruits of centuries of editorial labor to build one of the most powerful AI systems in history.
Encyclopedia Britannica and its subsidiary Merriam-Webster filed a copyright and trademark lawsuit against OpenAI on March 13, 2026, in the US District Court for the Southern District of New York. The complaint alleges that Microsoft-backed OpenAI used nearly 100,000 Britannica articles, including encyclopedia entries and Merriam-Webster dictionary definitions, to train its ChatGPT chatbot, all without permission or compensation.
What Britannica Claims
The lawsuit rests on two core legal pillars: copyright infringement and trademark violation. On the copyright front, Britannica argues that OpenAI’s systematic copying of its reference content went far beyond any reasonable interpretation of “fair use”, the legal doctrine that AI companies have consistently invoked as a shield against similar claims.
Critically, the lawsuit alleges that ChatGPT doesn’t merely draw on Britannica’s knowledge in a general sense, it can produce responses that are verbatim or near-verbatim reproductions of the publisher’s original entries. That detail is legally significant: it undercuts the argument that AI systems merely “learn from” text the way a human student would, rather than storing and regurgitating it.
The trademark angle is equally pointed. Britannica accuses OpenAI of harming its brand reputation by generating false AI “hallucinations” that incorrectly cite Britannica or Merriam-Webster as sources essentially, making up facts and attributing them to trusted reference brands. This, the company argues, constitutes trademark infringement by implying a permission and association that never existed.
The Traffic Cannibalization Problem
Perhaps the most commercially damaging claim in the complaint is the allegation of “web traffic cannibalization.” The argument goes like this: when a user asks ChatGPT a question about, say, the history of the Roman Empire or the definition of a word, the AI provides a polished summary, one derived, Britannica claims, from its own content, directly in the chat window. The user gets their answer without ever visiting Britannica.com or Merriam-Webster.com.
This dynamic strikes at the heart of the modern media economy. Publishers generate revenue through advertising and subscriptions that depend on page visits. If an AI model can serve up the essence of an article without the user ever clicking through, the economic model that funded the creation of that content in the first place collapses. Britannica’s lawsuit is, at its core, a demand to be compensated for the value it contributed to OpenAI’s most profitable product.
OpenAI Pushes Back
OpenAI’s response was swift and familiar. A company spokesperson stated that its models “empower innovation, and are trained on publicly available data and grounded in fair use.” This is the same defense the company has deployed in over a dozen other copyright lawsuits currently working their way through the US court system.
The legal theory is straightforward: training an AI model transforms text into statistical patterns, much as a human brain absorbs and synthesizes information from reading. Under this view, training is transformative and therefore qualifies as fair use, regardless of how much copyrighted text was consumed in the process.
Courts have not yet settled this question. The central fair use determination in the major AI copyright cases is not expected before summer 2026 at the earliest, according to analysts tracking the litigation landscape.
A Growing Legal Wave, Now at 91 Lawsuits
The Britannica filing brings the total number of copyright lawsuits filed against AI companies in the United States to 91, according to the ChatGPT Is Eating the World legal tracker. The case joins a sprawling multidistrict litigation (MDL) in the Southern District of New York, overseen by Judge Sidney Stein, which already consolidates more than a dozen suits brought by major news publishers, including the New York Times.
This is not Britannica’s first rodeo with AI litigation. The company filed a nearly identical lawsuit against Perplexity AI in September 2025, raising the same claims about AI-generated summaries diverting readers. That case remains pending.
The legal landscape is splitting in two. While publishers sue, others are cutting deals: News Corp signed a licensing agreement with Meta worth up to $50 million per year in March 2026, and UK publisher Reach struck a usage-based deal with Amazon for its Nova AI model. OpenAI itself has previously signed licensing agreements with the Associated Press and other outlets, but those came after the training data in question had already been used.
What Britannica Wants
Britannica is seeking two remedies from the court: an unspecified amount of monetary damages for the alleged infringement, and an injunction, a court order, blocking OpenAI from further use of its material in AI systems. The injunction request is notable: if granted, it could force OpenAI to retrain or modify its models, a technically complex and costly undertaking.
Legal analysts expect the case to be transferred to the existing AI copyright MDL and stayed pending that litigation’s outcome, potentially delaying a final resolution by years. The precedent being set elsewhere in the MDL, particularly around the fair use question, will likely determine the fate of Britannica’s claims as well.
The Bigger Picture
There is a certain irony to Britannica’s position at the center of this dispute. The encyclopedia was already declared dead once, by the internet. Its print edition ceased in 2012, and the company pivoted to digital, rebuilding itself as a premium online reference platform. It survived disruption by adapting.
Now it faces a second, more intimate disruption: an AI system that may have learned to do what Britannica does, partly from Britannica itself. The lawsuit is not just a legal filing, it is a philosophical argument about who owns knowledge, who should profit from it, and whether the companies building the AI revolution have an obligation to compensate those whose intellectual labor made it possible.
