Bezos-Backed Perplexity AI Accused of Free-Riding on Journalism

Journalists aren’t thrilled about being turned into free training fodder for Silicon Valley’s latest obsession. Now Japan’s Nikkei, owner of the Financial Times, and the Asahi Shimbun have filed a lawsuit accusing US-based Perplexity AI of industrial-scale plagiarism dressed up as “answer search.”

Nikkei and Asahi Take Aim at AI “Free Riding”

In a joint statement, the publishers accused Perplexity of “ongoing free riding” on journalism their reporters spent “immense time and effort” producing. Translation: stop strip-mining our work for free.

The companies want an injunction to force Perplexity to delete scraped content and halt the practice altogether. They’re also asking for damages of 2.2 billion yen each (around £11 million). Not exactly loose change for an AI startup that has yet to explain why “democracy” depends on its bots hoovering up newsroom copy.

Why This Matters for Media and Law

If unchecked, Nikkei and Asahi argue, Perplexity’s model risks “undermining the foundation of journalism” and—cue the lofty phrasing—“threatening democracy itself.”

Hyperbolic? Maybe. But they’re not wrong that the economics of news collapse if AI companies can simply ingest articles, answer user queries, and bypass paywalls.

This is not an isolated skirmish. It’s the latest in a cascade of lawsuits against AI companies accused of copyright theft. From the New York Times’ action against OpenAI, to a growing pile of authors and news outlets suing Anthropic, the AI industry is learning the hard way that “ask forgiveness, not permission” doesn’t always play in federal court.

Perplexity’s Position of Strength—For Now

Unlike Google, which at least pretends to funnel traffic back to publishers, Perplexity sidesteps the courtesy of links. Its model delivers polished answers by scraping live web content. That hasn’t stopped it from bagging deep-pocketed backers like Jeff Bezos and Nvidia.

Perplexity claims between 22 and 30 million monthly users. That kind of traffic doesn’t go unnoticed by lawyers, especially when most of it is built on someone else’s IP.

The Anthropic Twist

Meanwhile, across the Pacific, Anthropic has been wrangling its own legal mess. The company recently reached a class-wide settlement with authors in Anthropic v Bartz, pausing discovery while they iron out terms.

Judge William Alsup, (pictured) who has a history of schooling tech companies in copyright law, ruled earlier this year that using legally bought books for training was fair use. But storing more than seven million pirated ones? Not so much.

Reuters reports Alsup ordered a December trial to set damages, though the deal now looks set to short-circuit that showdown.

The Bigger Picture

The Nikkei/Asahi case drops into a legal landscape where courts are slowly drawing the line between “innovation” and “theft.” AI boosters love to describe these cases as quaint legacy industries lashing out at progress. The reality is starker however as newsrooms pay reporters; AI companies don’t. The result are diminished returns and tough times for media.

Strip away the rhetoric and the AI-based business model looks a lot like free-riding with a Silicon Valley gloss – to the major detriment of the media companies facing the rough going generated by digital change.

If Perplexity loses, it won’t just be a $44 million headache. It could set a precedent forcing AI companies to finally pay for the journalism, books, and creative work they’ve been gorging on. Until then, lawyers are the only winners—again.