The Search Engine Built on Stolen Content Faces a Legal Reckoning

Perplexity AI reached a $20 billion valuation in September 2025, making it one of the most valuable AI startups in the world. The company positioned itself as the future of search—an AI-powered answer engine that could synthesize information from across the web and deliver concise, sourced responses. The problem? Much of that "synthesis" turns out to be straightforward plagiarism, and the publishers whose content Perplexity scraped are now fighting back in court.

According to CNBC's reporting, The New York Times filed a lawsuit against Perplexity in December 2025, alleging the company "illegally copied and distributed" copyrighted content. This follows similar suits from the Chicago Tribune, News Corp (Wall Street Journal, Barron's, New York Post), Encyclopedia Britannica, Merriam-Webster, and international outlets like Nikkei and Asahi Shimbun.

The Scale of the Plagiarism Problem

This isn't a few edge cases or technical misunderstandings. According to TechCrunch's investigation and Skim AI's analysis, the evidence is damning:

• 48% Direct Plagiarism: AI plagiarism detection tool Copyleaks found that some Perplexity summaries contain nearly half the text directly rewritten from source articles—far beyond any reasonable definition of "fair use."

• Spoofed User Agents: According to analyses by Wired and Cloudflare, Perplexity uses undisclosed web crawlers with spoofed user-agent strings to scrape content from websites that explicitly prohibit such scraping.

• Robots.txt Ignored: Perplexity's crawlers reportedly ignore robots.txt files—the standard mechanism websites use to communicate what content can be crawled.

• No Attribution Revenue: Unlike traditional search engines that send traffic to publishers, Perplexity keeps users on its platform, depriving publishers of both traffic and advertising revenue.

"The suit accuses Perplexity of unlawfully scraping The Times' stories, videos, podcasts and other content to formulate responses to user queries. This isn't aggregation or summarization—it's content theft at scale." — TechCrunch Coverage of NYT Lawsuit

CEO's Dismissive Response: "Publishers Have Been Suing Tech for 100 Years"

Perhaps most telling is Perplexity's response to the mounting legal pressure. According to Wikipedia's documentation, CEO Aravind Srinivas claims Perplexity only "aggregates" rather than plagiarizes information. But the company's head of communication, Jesse Dwyer, offered an even more dismissive take:

"Publishers have been suing new tech companies for a hundred years, starting with radio, TV, the internet, social media and now AI. Fortunately it's never worked, or we'd all be talking about this by telegraph." — Jesse Dwyer, Perplexity Head of Communication

This cavalier attitude ignores a crucial distinction: radio, TV, and social media didn't wholesale copy newspaper articles and present them as original content. They created new content or provided platforms for user content. Perplexity's model is fundamentally different—it takes other people's work, repackages it slightly, and monetizes it without compensation.

The Growing List of Plaintiffs

According to Just Think AI's legal analysis, the lawsuits keep piling up:

• The New York Times (December 2025): The most high-profile plaintiff, with deep legal resources and a history of IP protection.

• Chicago Tribune (December 2025): Another major newspaper joining the fight.

• News Corp: Rupert Murdoch's empire, including Wall Street Journal, Barron's, and New York Post—formidable opponents with extensive legal teams.

• Encyclopedia Britannica & Merriam-Webster: Reference publishers whose entire business model depends on content IP.

• Nikkei & Asahi Shimbun: Japanese media giants, extending the legal challenge internationally.

• Reddit: Even user-generated content platforms are pushing back against unauthorized scraping.

The Revenue Sharing Attempt: Too Little, Too Late?

Facing legal pressure, Perplexity has attempted to make peace with publishers. According to Viinyx's coverage:

• Publishers' Program: Offers outlets like Gannett, TIME, and Fortune revenue sharing—but only to those who agree to participate.

• Comet Plus: Launched in August 2025, allocating 80% of its $5 monthly fee to participating publishers. But $4 per month split among all the publishers a user accesses is pocket change compared to the advertising revenue lost.

• Getty Images Deal: A multi-year licensing deal for images—acknowledging that at least some content requires payment.

These programs look more like legal strategy than genuine fairness. Publishers who don't participate still get scraped; they just don't get paid. And the revenue sharing amounts are trivial compared to the value Perplexity extracts.

The Technical Evidence of Bad Faith

According to Antlerzz's technical analysis, Perplexity's scraping practices suggest intentional evasion rather than accidental overreach:

• User-Agent Spoofing: Perplexity's crawlers disguise themselves as regular browsers to bypass crawler detection.

• Dynamic IP Rotation: The company rotates IP addresses to avoid rate limiting and blocking.

• JavaScript Rendering: Unlike simple scrapers, Perplexity renders full pages to access content protected behind JavaScript.

• Paywalled Content: Reports suggest Perplexity has accessed content behind paywalls, though the company disputes this.

The Precedent Problem

Perplexity's legal theory—that AI "transformation" of content constitutes fair use—hasn't been tested at this scale. According to JURIST's legal analysis, the outcomes will shape the entire AI industry:

• If Publishers Win: AI companies will need to license content, dramatically increasing costs and potentially making some business models unviable.

• If Perplexity Wins: Any AI company can scrape any content and claim "transformation," essentially making copyright irrelevant for text.

• Likely Outcome: Settlement or narrow rulings that create a patchwork of obligations, satisfying no one.

Why the $20 Billion Valuation Might Be a Mirage

Perplexity's valuation assumes continued growth without existential legal risk. But consider the scenarios:

• Injunction Risk: Courts could order Perplexity to stop using content from plaintiffs, gutting its knowledge base.

• Damages Liability: Statutory damages for willful infringement could reach billions across all plaintiffs.

• Licensing Costs: If forced to pay for content, margins collapse. Perplexity would need to either raise prices dramatically or accept losses.

• Advertiser Caution: Brands may avoid association with a company accused of content theft.

The User's Dilemma

For users who've come to rely on Perplexity, the ethical calculus is uncomfortable. Every query you run potentially uses content that was taken without permission. The convenience of AI search comes at a cost—not to you, but to the journalists, researchers, and publishers who created the underlying information.

The question isn't whether Perplexity is useful—it clearly is. The question is whether that usefulness justifies a business model built on taking other people's work without compensation. The courts will ultimately decide the legal question. The ethical question is one each user must answer for themselves.