Starting October 30, Meta will implement updated commercial terms of use that grant the company broad permissions to scrape content from websites linked to its business accounts. These accounts are used by companies and public figures to manage Facebook pages, Instagram profiles, WhatsApp channels, and run advertisements via a centralized dashboard. Because nearly every news organization maintains at least one such account, legal experts suggest these changes could provide Meta with a mechanism to harvest journalistic content to train its artificial intelligence models.
The image provides a visual reference for Meta, the company implementing the updated commercial terms for its business accounts.
The new language explicitly authorizes Meta—or entities acting on its behalf—to use tools like web scrapers, bots, and crawlers to access, index, cache, and analyze domains or URLs associated with the user’s products or services. The terms specify that this collected data can be used to "train" Meta products; in other parts of its service agreements, the company defines "AI at Meta" as one of those core products. Legal analysts note that while publishers may attempt to block crawlers using standard robots.txt files, the new commercial agreement may supersede those external instructions. Furthermore, the terms permit Meta to create "test credentials," which experts say could potentially allow the platform to bypass registration walls or paywalls to reach protected content.
Meta has stated that these updates consolidate previous crawling policies into one location and do not change how it crawls publishers in the future, noting that owners can still utilize robots.txt for management. However, critics point out that there is currently no option for business accounts to opt out of these terms without ceasing all commercial use of Meta's platforms. This move has been compared by some observers to an antitrust concept known as "tying," where access to essential advertising and social media infrastructure is contingent upon accepting broad data-sharing permissions. While any actual training of AI models remains subject to ongoing legal debates regarding fair use and copyright, the updated contract provides a direct pathway for Meta to claim permission from the very entities whose content it seeks to ingest.