Meta has revised its commercial terms of service, with the changes taking effect on October 30. The new language removes the ability for business accounts to decline the provision, instead urging firms to discontinue use of Meta services if they refuse to permit Meta’s automated agents to access their linked content. The amendment applies to all business accounts across the platform.
The revision is especially consequential for news organizations that currently rely on Meta platforms for audience reach. Those publishers have been negotiating licensing arrangements that could generate substantial payments for the use of their articles in AI training. Ongoing court disputes with another AI developer over whether such training constitutes fair use heighten the stakes, and Meta’s terms appear to grant it broad permission regardless of future rulings.
The updated clause authorizes Meta or any representative to employ automated software, including crawlers and bots, to index, cache, analyze, and retrieve content from domains and URLs supplied by the business. It also permits the creation of test credentials to access the same resources, with the stated purpose of supporting Meta’s products and services. Although the language does not explicitly reference artificial intelligence, the inclusion of the term “train” alongside references to automated scraping suggests an implied AI application.
Three legal scholars consulted by a media research outlet expressed uncertainty about how courts would treat the provisions. One expert noted that the terms could be invoked to argue that Meta is entitled to scrape news sites for AI training. Professor James Grimmelmann of Cornell Law School told the outlet that Meta could “train AI on any content from their sites” and might “deploy some of that content” in competition with the original publishers on its platforms.
A Meta spokesperson responded that content owners may still control access through standard robots.txt files. The company compared the new language to earlier agreements such as a seller contract that limited automated access to e-commerce listings, and a self-service advertising policy that allowed crawling only for ad generation. Critics argue that the current commercial terms introduce broader permissions, including the novel “test credentials” provision, which was not present in those prior documents.
The revision has prompted some publishers to reconsider their relationship with Meta’s social channels. The editor of an independent online magazine said he is weighing a departure after observing repeated scraping of his site. Nieman Lab reported that it is reaching out to affected news outlets to gauge whether the update could trigger a broader withdrawal from the platform.