New York Times launches lawsuit accusing Perplexity of “illegal” article copying

The New York Times has filed a wide-ranging lawsuit against Perplexity AI, alleging unlawful scraping and reproduction of millions of its articles, as reported by The Guardian. The case marks a significant escalation in the ongoing confrontation between major publishers and fast-growing AI platforms over the use of copyrighted journalism to power commercial models.

Key Points from The Guardian’s coverage

• The Times alleges Perplexity has copied and distributed its journalism “en masse” without permission and has violated its trademarks by generating fabricated content and presenting it alongside the newspaper’s branding.

• The complaint states that Perplexity’s business model depends on scraping a broad range of content, including paywalled material, to train and run its generative AI tools.

• Perplexity is facing multiple legal challenges, including suits from Dow Jones, the New York Post, the Chicago Tribune, Merriam-Webster, Encyclopedia Britannica and Reddit, as well as accusations from Cloudflare that it hid its web-crawling activity.

• The company has raised about $1.5bn over three years and was most recently valued at $20bn, with backing from investors such as Nvidia and Jeff Bezos.

• Amazon has also filed a separate lawsuit alleging Perplexity covertly accessed Amazon user accounts through its AI shopping agent. Perplexity has denied wrongdoing in all cases.

Analysis

The Times’ latest filing underscores how rapidly the legal landscape is hardening around the commercial use of publisher content in AI systems. Although litigation between news organisations and tech companies is not new, the breadth of claims now facing Perplexity suggests an industry reaching a breaking point. Publishers are no longer treating scraping as a background cost of doing business but as an existential threat to the value of their intellectual property.

For media companies, the case highlights two converging pressures. First, the shift from early-stage experimentation to full commercial deployment of AI agents is creating new incentives for platforms to ingest copyrighted content at scale. Second, the perceived opacity of some AI models’ data practices is damaging trust at a moment when publishers are increasingly seeking structured licensing agreements to safeguard both revenue and reputation.

From my perspective, this dispute will likely accelerate moves toward clearer contractual frameworks for training data, with publishers asserting more control over how their journalism is used. At the same time, regulators may feel emboldened to intervene where they see systemic risk to competition or consumer trust.

Two credible scenarios emerge from here. One is a negotiated settlement that helps set industry norms for licensing and attribution, which could stabilise relations between publishers and AI firms. The other is a more fractious outcome in which divergent court rulings create a patchwork of precedents, leaving both media and AI companies to navigate prolonged uncertainty. Either direction will have significant implications for how news content is valued, protected and commercialised in the next phase of AI adoption.