Meta’s updated Commercial Terms, due to take effect on 30 October 2026, would permit it to crawl websites associated with business accounts and use their content to train its products, Nieman Lab reported on 15 October. The permissions would extend to news publishers that use Meta’s business accounts, potentially giving the company a route to their websites without a separate licensing agreement.
Key points
- The terms cover domains and webpages associated with a business user’s products or services, and explicitly list training among permitted uses.
- Meta says the update consolidates existing language and will not change its crawling practices. Legal experts question how the terms would interact with website blocking instructions.
- Meta already has AI licensing deals with CNN, Fox News, USA Today and People Inc.
Meta’s terms reach business-linked websites
The reported wording permits Meta and parties it appoints to deploy automated software on domains, webpages and URLs linked to a business user’s products or services. It also covers sites made available through use of Meta products. The listed activities include indexing, collecting and periodically retrieving content, as well as creating test credentials to access websites.
Meta may use content gathered from those sites to develop, research and train products covered by the Commercial Terms, according to the language published by Nieman Lab. The business accounts affected can be used to manage Facebook pages, Messenger accounts, Instagram profiles and WhatsApp channels. They also cover activities including advertising, marketing messages and use of Meta’s business dashboard.
Business users who reject the update are told to stop using Meta products for commercial purposes. Nieman Lab reported that the terms offered business accounts no advance opt-out from the update. That condition matters to publishers using those accounts to distribute stories or manage their presence on Meta’s platforms.
Earlier Meta terms covered advertising URLs
Meta’s earlier terms for generative AI in advertising allowed crawling of URLs users supplied in connection with ads. The new Commercial Terms cover a broader category of sites associated with business users and expressly include training, NeoTeo reported. The updated provision sets out permission to collect content rather than reporting that Meta has crawled a particular publisher’s site.
Meta has also made AI licensing agreements with CNN, Fox News, USA Today and People Inc., The Verge reported on 16 October. Those deals provide a commercial comparison for the access contemplated by the new terms. Matt Topic, a media and intellectual property attorney at Loevy & Loevy, said the provision could give Meta access to material likely to cost many millions of dollars in a licensing market. He said it was uncertain whether a court would enforce the provision.
The Verge reported that Meta launched Meta One in September, a subscription charging businesses and creators for links they place in posts and reels. The proposed crawling permission concerns websites outside Meta’s platforms, including sites to which business accounts may direct readers through those links.
Meta and lawyers differ on robots.txt
A Meta public affairs spokesperson told Nieman Lab that the update brings together crawling language previously spread across other terms. The spokesperson said Meta’s crawling practices would not change and content owners could continue to manage crawler access through robots.txt, a file that gives automated visitors instructions about parts of a website.
Legal experts consulted by Nieman Lab said the wording could allow Meta to argue that its Commercial Terms take precedence over website blocking instructions. James Grimmelmann, a professor of digital and information law at Cornell Law School, said the permission could extend to training on news-site content. On its broadest interpretation, he said, Meta could use the provision to build a comprehensive archive of that content.
The test-credentials clause presents another possible route to website content. Grimmelmann said it could technically allow Meta to register as a user to pass a publisher’s registration wall. Legal experts said any training using content collected under the terms would still depend on unresolved questions about copyright and fair use.
Five large book publishers filed a class-action suit in May alleging that Meta pirated millions of books to train its Llama models.
