Meta Updates Terms to Allow AI Training on Business Users' Web Content

Meta has updated its commercial terms of use to explicitly authorize the scraping and use of web content associated with business accounts on Facebook, Instagram, and WhatsApp for the development and training of its AI products. The changes, which take effect on October 30, 2026, state that users authorize Meta to use automated software to access, index, and crawl domains and URLs linked to their products or services. Meta may then use this content to "provide, develop, improve, research, train, and maintain" its services.

No Opt-Out and Paywall Bypass Concerns

Legal experts and publishers note that there is no mechanism for business accounts to opt out of these new terms; continued use of Meta’s business tools requires acceptance of the new scraping permissions. The terms also allow Meta to create "test credentials," which legal analysts suggest could allow the company to bypass registration walls or paywalls to access content. While Meta claims the update merely consolidates existing scattered language and that robots.txt files will still be respected, legal scholars argue the new commercial terms could be interpreted as superseding standard web blocking instructions.

Legal and Industry Implications

The move has drawn criticism from legal experts who describe it as a form of "tying," where access to Meta’s business platforms is contingent on granting broad data collection rights. This parallels recent changes by Google regarding AI Overviews. The update comes as Meta faces ongoing litigation regarding AI training data, including a class-action suit alleging the company pirated books to train its Llama models. While Meta states that robots.txt files remain a valid method to restrict access, the new terms create legal ambiguity regarding whether commercial agreements supersede technical blocking mechanisms like robots.txt.