On November 2, 2026, The New York Times joined the New York Daily News, the Chicago Tribune and other US publishers in suing Brave Software, alleging the browser maker scraped copyrighted news content to train its AI features and gave users a route around paywalls. The cases push AI copyright disputes past model training and into the browser, the layer that decides how readers actually reach a story. If the claim holds, built-in summaries and on-device answering will carry new licensing costs,
Reuters broke the story on November 2, 2026, reporting that the New York Daily News, the Chicago Tribune and other outlets had lodged complaints against Brave Software, with The New York Times joining as a co-plaintiff across a cluster of related filings. The complaints make two central allegations: that Brave scraped news content without authorisation to train its Leo AI assistant and its page Summarizer, and that the browser's design lets readers sidestep publisher paywalls to reach article text that would otherwise require a subscription.
Brave's business model ties the two claims together. The company runs its own search index, which it says spans more than 30 billion pages, and licenses results to third parties through the Brave Search API, while Leo and Summarizer generate answers on the user's side and reduce clicks to the original site. The plaintiffs seek injunctions,
Copyright litigation usually trails a generation behind product design. When The New York Times sued OpenAI and Microsoft in December 2023, the question was whether training had reproduced protected works. When it sued Perplexity in August 2025, the focus shifted to generated answers replacing the act of reading, but the defendants were still AI search companies. Browsers stayed on the sidelines as presumed neutral tools until they grew summarisation and chat features of their own.
The technical ground has shifted too. In July 2025 Cloudflare began blocking known AI crawlers by default, and in September 2025 Reddit, Yahoo, Ziff Davis and others backed the Really Simple Licensing standard,
Liu Gong's take: these suits will not end with Brave switching off its crawler within twelve months. The likelier landing zone is a licensing deal paired with output-side constraints, and within six months a second major browser vendor will most likely sign a citation or content-licensing agreement with publishers, turning compliance cost into a competitive moat.
From a technical standpoint, the plaintiffs' tools are blunt. robots.txt and RSL live on the server, while Brave's summaries and Leo conversations run on the user's device, with fetch requests buried inside ordinary reader sessions that a server cannot easily separate from crawlers. The enforceable boundary sits on the output side: snippet length, whether a clickable source link is mandatory, and how much of an article may be quoted. Pew's 2025 figures put a price on that line, with 8% of users clicking through when a summary appears against 15% when none does. If a court accepts that summarisation substitutes for reading and damages the market, browsers will have to concede at the rendering layer rather than the crawling layer.
On the business side, the size gap says this is not primarily a damages play. Brave reports roughly 100 million monthly actives and earns from search ads, Search API licensing and subscriptions, an order of magnitude away from OpenAI, so even a plaintiff victory yields limited recovery. The real currency is precedent and leverage, making in-browser summaries a paid licence category. Search API and Leo are precisely what distinguishes Brave from Chrome and Edge; a forced retreat would wound the product story.
For the industry, the copyright fight has moved from training to retrieval to distribution. In industrial inspection, every defect image used to train a model traces back to a production batch and a licence contract, so a dispute can be pinned to a specific link in the chain. Publishing lacks that chain and is rebuilding it one lawsuit at a time. Liu Gong predicts that before the first half of 2027, mainstream browsers will ship mandatory citation and click-through as a default setting, a compliance move that doubles as bargaining currency.
- The New York Times reported about 11 million digital subscriptions in the second quarter of 2025, with the paywall central to that revenue; Brave reports roughly 100 million monthly active users, so the two sides are far from symmetrical.
- Pew Research (2025): when an AI summary appears on a results page, 8% of users click an external link, versus 15% on pages without one, close to a halving.
- Under US copyright law, statutory damages for wilful infringement reach $150,000 per work; in the Times' suit against OpenAI, outside estimates of exposure have run into the billions of dollars.
- Brave says its independent index covers more than 30 billion pages and licenses results through an
For the industry, the near term of three months should bring more publisher filings and prompt browser and AI assistant vendors to shorten summaries and add source links. Within a year, content licensing should move from bespoke deals to a default expectation, with machine-readable licences such as RSL likely to be read by browsers out of the box. Smaller browsers will find compliance cost harder to absorb and drift toward platform owners or search API suppliers, concentrating the market further.
For ordinary users, the next three months may bring shorter summaries, more jump links and stricter bot detection on some sites, producing occasional walls where an article will not open. Within a
Questions worth exploring:
- Whether client-side fetching and summarising fall under server-side robots conventions, and where the technical line should sit
- How news licensing gets priced: by crawl volume, by citation count, or by share of subscription conversions
- Survival paths and consolidation prospects for smaller browsers as compliance costs rise
Further reading: 2026 Frontier LLM Comparison: A Practical Guide — a durable, evergreen reference. Bookmark it.
This item comes from The Verge AI (published 2026-11-02 17:51:19). Wenfeng Radar provides a Chinese/English summary and commentary on overseas AI developments; the original content is owned by its author.
Daily aggregation of overseas AI developments with in-depth analysis. Bookmark us to never miss an important signal; back to homepage.