The New York Times v. Brave: Scraping Paywalled News to Feed AI Companies Constitutes Copyright Infringement

Recently, The New York Times and the New York Daily News, among other publishers, separately filed lawsuits in the U.S. District Court for the Southern District of New York (Manhattan federal court) against Brave Software, the operator of the privacy-focused browser Brave. Breaking away from the traditional disputes over AI companies directly scraping content, this case targets a browser vendor as an intermediary in the copyright infringement chain, offering a new approach to copyright disputes over AI training data.

Factually, the plaintiffs raise several allegations in their complaints. First, Brave allegedly uses undisclosed crawlers to systematically scrape copyright-protected news content, building an index of over 30 billion web pages, without providing a compliant mechanism for publishers to block such access. "Undisclosed crawlers" refer to scraping programs that do not openly identify themselves, often masquerading as ordinary user visits and failing to comply with the robots protocol, making it difficult for content owners to identify and block them. Second, Brave allegedly leverages browser features to help users bypass publisher paywalls and access articles that would otherwise require a subscription. Third, Brave allegedly supplies and sells the scraped news content to AI companies such as Anthropic, Perplexity, and xAI for large language model training (note: these AI companies are not defendants in this case; the plaintiffs merely identify them as counterparties in transactions as a factual matter, rather than as targets of liability).

As for legal claims, The New York Times case asserts three copyright infringement theories—direct infringement, contributory infringement, and vicarious infringement—while also asserting claims under the Digital Millennium Copyright Act (DMCA) for circumvention of technological protection measures and a claim for unjust enrichment. The relief sought includes permanent injunctions, statutory damages, and disgorgement of infringing profits. The New York Daily News and other media outlets simultaneously filed a separate but highly similar complaint in the same court.

The core of this case is not simply whether crawler scraping constitutes fair use, but rather a novel liability model: whether a browser tool provider can be held liable as a copyright intermediary for providing paywall-circumvention features and packaging scraped content for sale to AI companies. In prior copyright lawsuits of this kind, plaintiffs mostly sued the AI companies training the models directly, with the dispute centered on "whether the training itself constitutes fair use." The plaintiffs' logic in this case is to disaggregate the chain, arguing that Brave is the scraper, distributor, and monetizing intermediary of the content, tying together three acts—tool functionality, crawler indexing, and B2B sales—and layering DMCA anti-circumvention provisions on top, creating evidentiary burdens and defense space that differ markedly from prior cases.

Two points merit particular attention. First, a paywall constitutes a technological protection measure under copyright law; the DMCA prohibits circumvention of such measures and also prohibits the manufacture or provision of circumvention tools. If Brave's paywall-bypass feature is found to be a tool specifically designed to circumvent publisher access controls, that claim could stand independently of the reproduction infringement claims. Second, a distinction must be drawn between "mere search engine indexing" and "selling access to indexed content to third-party AI companies." Traditional search snippet models still have room to assert fair use, but if the business model involves selling access to full articles, the scope for that defense is substantially narrowed.