News Corp labels some AI companies ‘crass kleptomaniacs’
News Corp has publicly attacked some AI companies for using its content without payment, with CEO Robert Thomson branding certain firms "crass kleptomaniacs" during the group's latest earnings announcement. He singled out browser maker Brave for particular criticism, accusing it of using masked web crawlers to scrape News Corp articles and repackaging them into near-verbatim copies sold to enterprise customers, while praising OpenAI and Meta, which have paid licensing deals with News Corp, as "principled" and "honorable" partners. The remarks underline the escalating legal and commercial battle between traditional publishers and AI firms over the unauthorised use of copyrighted material to train and power AI systems.
Thomson said News Corp was pursuing a "woo and sue" strategy, warning that companies using content pilfered by such AI firms are "in possession of stolen goods" and that further litigation is likely. He also highlighted the strong performance of the company's Digital Real Estate Services unit, including realtor.com, which grew revenue by 12 percent to $1.8 billion, as part of a wider full-year revenue increase of 7 percent to $9.03 billion; specific revenue from the OpenAI and Meta deals was not disclosed. The comments come as authors and other content creators pursue separate lawsuits against AI developers, including a $1.5 billion settlement Anthropic agreed to pay in a related case.
- News Corp's CEO called some AI firms "crass kleptomaniacs" over content scraping.
- Brave was accused of stealing and reselling News Corp's articles.
- OpenAI and Meta praised as paying, "principled" content partners.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Content creators and publishers invest substantial resources in journalism, research and entertainment, and copyright law exists to protect and fairly compensate this intellectual labour. If AI companies can freely use copyrighted works for commercial gain, it undermines the economic incentive to produce quality content and extracts value unfairly. Major AI firms like OpenAI and Meta have demonstrated that licensing agreements are commercially viable, suggesting principled competitors can operate profitably whilst compensating creators fairly.
The case against
Training AI systems on publicly available text is fundamentally transformative—the systems learn statistical patterns rather than reproducing original works, much as a scholar learns from reading many books without licensing each one. Copyright's fair use doctrine traditionally permits such non-consumptive research use, and requiring explicit licensing for every training document would be operationally impractical and could entrench incumbent technology companies by making AI development prohibitively expensive. Society benefits when knowledge is accessible and AI systems can learn from diverse sources, similar to how search engines and academic institutions have historically operated.