Encyclopedia Britannica Sues OpenAI: The Content Mining War Reaches Historic Proportions

Watching the Encyclopedia Britannica—a name that evokes the smell of old paper, leather-bound books arranged on wooden shelves, and librarians who were familiar with every Dewey decimal—enter a federal courtroom in Manhattan and file a lawsuit against one of the most potent artificial intelligence firms on the planet is almost poetic. It’s the kind of moment that simultaneously seems odd and inevitable. The collision of the old and new worlds—not in an abstract essay, but in a legal document bearing actual dates and monetary amounts.

Earlier this month, Britannica and its Merriam-Webster subsidiary filed a lawsuit against OpenAI, claiming that the ChatGPT developer had unlawfully copied almost 100,000 of its articles, dictionary entries, and reference materials.

The complaint’s wording is remarkably acerbic for a legal document. The term “cannibalized” appears, which is unusual for filings pertaining to intellectual property. It alludes to something ravenous and predatory. In a sense, Britannica appears to think that’s what’s taking place.

Britannica withstood the challenges of the digital era for many years. Wikipedia, Google, and the gradual demise of printed reference books—they all changed, went online, and persisted. It’s difficult to avoid thinking that AI might pose a completely different threat as this most recent chapter develops. Not because the technology is more intelligent, but rather because it silently absorbs the content and returns it to users in paraphrased sections, frequently without the reader ever clicking through to the original source.

That is the main point of contention for Britannica. The company claims that OpenAI trained ChatGPT using more than just its encyclopedia. It created a product that now rivals Britannica for the same visitors to its website. Traffic decreases. Slides with ad revenue. Interest in subscriptions declines. In the meantime, OpenAI’s valuation has skyrocketed above $700 billion—a sum that is nearly meaningless. One side created the library, and the other discovered a way to serve the books without ever giving credit, so it seems as though this asymmetry is the true story.

For its part, OpenAI has reacted as it typically does. According to a spokesperson, the models are based on fair use and trained on publicly accessible data. From the authors’ lawsuits to The New York Times’ lawsuit, the company has relied on the same defense in every case of a similar nature. The courts have not yet reached a decision regarding the viability of that defense. The earlier Anthropic settlement, which awarded $1.5 billion to authors whose books were pirated, suggested that judges are prepared to draw some boundaries, even though they are still debating where those boundaries should be.

In addition, Britannica filed a trademark claim, citing ChatGPT’s practice of misattributing information to the encyclopedia due to hallucinations. It feels almost intimate. Having a chatbot fabricate facts and put the Britannica name on them is especially offensive for a company that has spent 250 years establishing its reputation for accuracy.

Right now, there’s a more general change in attitude regarding AI. The initial awe has subsided. Protests are taking place at data centers. The cost of electricity is rising. And one by one, the individuals whose labor silently powered these machines are appearing in court. Perhaps not the largest of them is Britannica’s suit. However, the moment the encyclopedia itself decided it had read enough may be the one that future historians cite.