NYT sues Brave over AI scraping and its Search API

The New York Times sued Brave Software in federal court in Manhattan on November 2, alleging Brave scraped its journalism and sold access to AI companies through the Brave Search API. Hours later, the publishers of the New York Daily News, Chicago Tribune and Orlando Sentinel filed a similar suit. If your AI agent grounds its answers in a third-party search API, this is the lawsuit that asks where that data came from.

What actually happened

Both complaints target the Brave Index, the 30-billion-plus page collection behind Brave Search, according to TNW, which reviewed the Alden papers' complaint. The core allegations:

• An unnamed crawler. The publishers say Brave never disclosed the name of the bot that collects pages, so sites can't block it by name in robots.txt.
• Resale to AI companies. Brave sells index access through its Search API. Both complaints name the same five customers: Anthropic, Perplexity, Mistral, xAI and Cohere.
• Ask Brave. The AI answer feature allegedly reproduces passages word for word and summarizes articles in detail.
• Paywall blocker. A browser setting the papers say lets readers into subscription content without logging in.

The Alden papers bring copyright, DMCA circumvention and unjust-enrichment claims. They say more than two million registered articles are affected and want damages and a permanent injunction. A Times spokesperson told Reuters the company had repeatedly asked Brave to stop. These are allegations, not findings; Brave's response was not in the coverage we reviewed. Brave already has a separate fight with News Corp over the same indexing practices.

Why the Brave lawsuit matters for your business

Your search provider is part of your supply chain. Plenty of agents, RAG pipelines and "search the web" tools call a search API under the hood. If a court enjoins how that index was built, the API you depend on can change terms, drop sources or get pricier overnight. Know which provider your stack calls, including the ones buried inside a framework default.

Keep the search layer swappable. Put web search behind one function in your code, not scattered SDK calls. Switching from one provider to another should be a config change and a test run, not a rewrite.

Check your own front door. The complaint hinges on a crawler nobody could name. Look at your server logs and your robots.txt. If you care who trains on your product pages and blog, block the agents you can name and watch for the traffic you can't.

Sources: TNW, heise online.