Crass Kleptomaniacs: News Corp Goes to War With AI Companies Over Stolen Content

News Corp CEO Robert Thomson had some choice words for AI companies that train on copyrighted content without paying: “crass kleptomaniacs,” “pilferers,” and producers of “stolen goods.” As The Register reports, the tirade came during the company’s earnings call, where Thomson made clear that the Murdoch empire is pursuing a dual strategy: “woo and sue.” Pay up (as OpenAI and Meta have done), and you are “principled” and “honorable.” Scrape without permission, and brace for litigation.

The specific target of Thomson’s ire was Brave, the browser company he accused of deploying “masked web crawlers” to ingest copyrighted articles at scale and repackage them for enterprise customers. But the broader indictment applies to any AI company whose training data includes copyrighted work acquired without consent or compensation.

This is not a fight between David and Goliath. News Corp is a $9 billion media conglomerate with the legal resources to enforce its claims. The real significance of this story is what it signals for individual creators. As I discussed in Anthropic’s $1.5 billion settlement with authors, the precedent being set here applies to everyone who creates content—not just multinational publishers. The Perplexity incident, where Reddit caught an AI company red-handed scraping hidden content, shows the pattern is widespread. And the Wikimedia Foundation’s paid licensing deals with Microsoft, Meta, and Amazon suggest the emerging model: consent and compensation, not free scraping.

The question Thomson’s remarks force is uncomfortable but necessary: if a $9 billion publisher has to fight this hard to get paid for its content, what chance does an individual writer or artist have? The answer, for now, is collective action—lawsuits, settlements, and licensing deals that establish the principle that AI training data is not a free resource.

Why this matters now: The “woo and sue” era is just beginning. As more publishers and creators demand compensation, the cost of training AI on copyrighted content will become a defining economic issue for the industry. For Cogtropolis readers, this is the front line of individual economic agency in the AI age.