Perplexity — the buzzy AI “answer engine” that promises to let people skip the links and get straight to a neat summary — was hit with another heavyweight lawsuit this week. On Sept. 10, 2025, Encyclopædia Britannica and Merriam-Webster (Britannica owns Merriam-Webster) filed a federal complaint in Manhattan accusing Perplexity of copying, scraping and otherwise misusing their content to fuel the company’s AI answers. The case is the latest in a growing stack of suits that pit traditional publishers against startups building on the public web.

What the complaint says

Britannica’s lawyers lay out a simple story: Perplexity’s system uses a crawler (the complaint calls it “PerplexityBot”) to harvest material from Britannica and Merriam-Webster’s sites; the company then stores and uses that material to produce the short, chatty answers users see in Perplexity’s interface. According to the filing, sometimes those answers are near-verbatim reproductions of Britannica or Merriam-Webster text, and sometimes Perplexity attaches the Britannica/Merriam-Webster brand to answers that are incomplete or simply wrong — a trademark injury claim wrapped around the copyright claim.

The complaint even includes side-by-side screenshots meant to show the point: a Perplexity result reproduced Merriam-Webster’s definition of a word so closely that the plaintiffs call it plagiarism. That concrete example is a centerpiece of the filing and the kind of visual evidence judges find compelling.

What the plaintiffs are asking for

Britannica and Merriam-Webster want the court to stop Perplexity from continuing the alleged conduct — they seek injunctions to bar copying and misuse of their trademarks — and unspecified monetary damages. The complaint frames the case as both traditional copyright infringement and a reputational harm claim: a respected reference brand being used to prop up machine-generated content that the brands didn’t create or vet.

Why this matters beyond dictionaries

At first glance, a dictionary and an encyclopedia suing an AI startup might look niche. But the dispute cuts to the heart of how so-called Retrieval-Augmented Generation (RAG) systems work. RAG-style tools pull documents from the live web as part of the answer-generation pipeline; if those documents are stored, indexed or echoed back in user-facing answers without permission, publishers argue their traffic and revenue are being redirected into the AI product rather than to the original site. That’s the core economic gripe most publishers have with modern AI search: the box that answers your question may be replacing the link that used to send you to a newsroom or reference page.

Perplexity’s place in the broader fight

This is not Perplexity’s first courtroom rodeo. Over the last year, the startup has faced letters and lawsuits from news organizations — including Dow Jones / The Wall Street Journal and The New York Post — and public complaints from outlets like Forbes and The New York Times. Publishers have accused Perplexity of “stealth crawling” sites — bypassing publisher crawler-blocking tools — and of presenting reproduced content without adequate attribution or compensation. That pattern helps explain why Britannica’s suit landed in Manhattan: it’s one node in a web of disputes over who controls the value of internet content in an age of generative AI.

The company at the center of the storm

Perplexity has positioned itself as a modern alternative to Google’s “ten blue links.” Investors have poured large sums into the company: past funding rounds and reporting show support from heavy hitters, including Jeff Bezos and NVIDIA, and the startup has aggressively expanded features and ambitions — even experimenting with commerce and its own browser projects. That growth has put Perplexity squarely in competition with established search and media firms, at the same time those firms are trying to protect their content.

Publishers, partners and the messy middle

Not every publisher has fought Perplexity. The company has rolled out ad-revenue and publisher programs that some outlets have joined — Time and the Los Angeles Times have publicly participated in revenue sharing, for instance. Meanwhile, other reference sites have leaned into Perplexity: a small but notable example is World History Encyclopedia launching a Perplexity-powered chatbot to help users search its database of academic material. That split — licensing vs litigation — is a live signal that the market hasn’t agreed on what fair use, compensation or attribution should look like for generative search.

What the lawsuit might change (or not)

There are a few plausible outcomes. The case could push Perplexity toward licensing deals and stricter crawling behavior — publishers hope to see paid relationships or product changes that route users back to original pages. Alternatively, Perplexity could defend a model of transformation and fair use, arguing that summarizing public web content is a tech evolution of search. Whatever happens, judges are being forced to apply decades-old copyright and trademark law to a new kind of internet product, and their rulings will ripple through the AI industry.

The bigger picture

This lawsuit is another chapter in a broader reckoning: publishers are no longer passive data sources for AI builders. They are litigating, negotiating and experimenting with programmatic deals — and the law will likely determine whether the next wave of internet search leans on licensed, paid content or on unlicensed “summaries” that risk hollowing out the original creators’ traffic and revenue. For readers and researchers, who pay for high-quality, human-edited knowledge, it matters how reliable that knowledge will be in a world of instant, AI-generated answers.