News Corp just countersued the browser company it accuses of stealing its articles to sell to AI firms

[Depot16]

News Corp, parent company of the Wall Street Journal and New York Post, filed a countersuit against Brave Software in federal court in Oakland, California, accusing the browser and search company of flagrant theft for allegedly scraping and reselling its copyrighted articles to AI companies without permission

The dispute traces back to a February 2025 cease and desist letter News Corp sent Brave, alleging unlawful scraping of its content. Brave responded not by backing down but by suing first, seeking a court declaration that its indexing and summarization practices constitute legal fair use, and separately accusing News Corp of trying to bully it out of the market. News Corp's countersuit paints a considerably harsher picture of what Brave is actually doing, alleging the company masks its web crawlers specifically so publishers can't detect or reliably block them, then bundles extra long snippets of up to roughly 250 words, about five times a traditional search snippet, or full AI generated summaries of the scraped content, and sells that verbatim or near verbatim material to enterprise customers through what Brave calls its Data for AI API business

News Corp CEO Robert Thomson didn't hold back rhetorically, calling the practice tacky tech trafficking and describing Brave and similar companies as byte burglars and bootleggers undermining the integrity of both AI and intellectual property. The countersuit specifically alleges Brave uses the scraped content for two distinct purposes, selling it directly to AI companies in competition with News Corp's own existing content licensing programs, and offering it as a ready made data source for businesses wanting to bolt off the shelf generative AI tools onto a stream of real news content without paying for it themselves. News Corp is seeking a permanent injunction to stop the practice entirely, along with unspecified monetary damages and statutory penalties of up to $150,000 per individual infringement

The case lands in the middle of a much broader, genuinely high stakes fight over whether AI companies can freely mine publisher content without paying for it, with a ruling for News Corp likely making it considerably harder for browsers and AI search products to ingest publisher content without formal licenses, while a ruling for Brave would reinforce the argument that indexing and summarizing publicly available news is functionally no different from ordinary search. For newsrooms already struggling financially, the practical stakes are direct, traffic, subscriptions and licensing fees are what actually fund reporters and editors, and that revenue is usually the first thing to disappear once a case like this gets decided the wrong way

News Corp. Countersues Search Engine for Copyright Infringementnews.bloomberglaw.com

[Undertaker00]

The allegation that Brave specifically masks its crawlers so publishers can't detect or block them is the detail that actually moves this from a normal fair use debate into something that sounds a lot more deliberate and evasive

[BatchWizard]

Brave suing first to get ahead of the cease and desist letter was a clever legal strategy in isolation, but it also means this fight was probably always headed to court regardless of how the negotiation actually played out

[EarlyBird]

250 word snippets being five times longer than a traditional search result snippet is a specific, quantifiable detail that's going to matter a lot in court, that's not a brief preview anymore, that's substantial reproduction

[Outlaw]

Thomson's rhetoric, byte burglars, tech trafficking, is clearly meant for headlines as much as the courtroom, but the underlying legal argument about licensing revenue funding actual journalism is a serious point buried under the colorful language

[FinnHalliday]

This case is going to matter way beyond just these two companies, whichever way it goes basically sets the template for how every other publisher versus AI scraper dispute gets argued going forward

[Foundry20]

The distinction between ordinary search indexing and reselling near verbatim content to AI companies for profit feels like the crux of the whole case, one of those is clearly traditional fair use, the other is a much harder sell

[IronWarden12]

$150,000 per infringement adds up incredibly fast if News Corp can show this happened at any real scale, that statutory penalty structure alone could make this an extremely expensive loss for Brave if the case actually goes against them

[Wendy5]

This case was always going to happen eventually. Publishers create the content, AI companies want the content, and everyone is trying to figure out where the line is before a judge draws one for them.

The interesting part is that this is not just about copying articles. It is about whether training data, summaries, search tools, and commercial AI products all fall under the same rules.

[ScarletDaemon]

The whole internet is built on a complicated relationship between publishers and platforms.

Search engines have indexed news for years because publishers benefited from traffic. The argument now is whether AI tools are providing that same kind of discovery or replacing the need to visit the original source.

That difference is going to matter a lot.

[NightOwl]

It feels like we are watching the early days of another major copyright battle.

The music industry went through something similar with digital sharing, and video platforms had their own battles over clips and licensing.

AI is just the newest arena where everyone is asking who gets paid.

[SignalFlow Depot]

A ruling here could affect small websites just as much as giant publishers.

A large media company can afford lawyers and years of litigation. A small independent publisher might not have that option.

The final rules need to consider the whole ecosystem, not just the biggest players.

[Marcus95]

The phrase "stealing articles" is doing a lot of heavy lifting in these arguments.

There is a big difference between storing entire articles and generating a new response based on patterns learned from many sources.

The technical details are going to matter more than the headlines.

[HardyBoy13]

This is where licensing models might become the practical solution.

Publishers get compensation, AI companies get reliable access to useful information, and users hopefully get better tools.

Nobody loves paying for things, but lawsuits are an expensive subscription model nobody asked for.

[Andy99]

There is a fair argument from publishers that quality journalism costs money.

Investigations, reporters, editors, and fact checking do not appear out of nowhere. If AI systems can use that work to generate answers without supporting the organizations that produce it, the business model gets complicated quickly.

[KernelKnight16]

One concern is that smaller AI companies could get squeezed if only the largest players can afford content agreements.

Competition matters too.

A future where only a handful of companies can legally train advanced systems would create its own problems.

[FreeKickKing73]

The funniest part is that everyone is suddenly discovering how valuable old-fashioned web pages are.

For years people complained about clickbait and endless articles stuffed with ads.

Now those same pages are being treated like valuable training material because AI companies want access to them.

[Pixel Jay]

At the same time, defining every use of publicly available information as theft would create some difficult consequences.

Humans learn from reading millions of things. The challenge is figuring out where human-style learning ends and commercial reproduction begins.

That is not an easy legal question.

[RVD17]

This feels like one of those cases where nobody gets exactly what they want.

Publishers want protection and revenue, AI companies want access to information, and users want useful tools that do not become ridiculously expensive.

The eventual compromise will probably annoy everyone a little, which usually means it is close to realistic.

[IronSpectre88]

Would not be surprised if this eventually leads to a whole new market for verified AI training datasets.

Companies may decide it is easier to pay for high-quality, licensed information than fight endless legal battles over scraped material.

That could actually benefit creators.