Skip to content
NewsRegulation

Meta deletes AI-training clause from its commercial terms after legal pushback

· by Pondero Newsdesk

The short version

Meta removed language from its updated commercial terms of use that legal experts said could let it scrape and train AI on any publisher site tied to a business account, two days after the clause surfaced.

Meta deletes AI-training clause from its commercial terms after legal pushback

Meta removed a clause from its updated commercial terms of use on October 17 that had told business-account holders Meta could "use any content from" their "domains, web pages, or URLs to provide, develop, improve, research, train, and maintain" its products, two days after legal experts flagged the wording as a possible route around publishers' robots.txt blocks, per The Verge.

What happened

The language appeared in a revision to Meta's commercial terms that was set to take effect October 30 and covered any business promoting content across Facebook, Instagram, WhatsApp, or Messenger, according to The Verge's original report, which cited Nieman Lab's reporting on the clause. Cornell digital-law professor James Grimmelmann told Nieman Lab, as quoted by The Verge, that the terms "clearly represents permission to Meta to train AI on any content from their sites and possibly to deploy some of that content, even in competition with the publishers themselves." Meta communications VP Andy Stone initially pushed back on X, saying the update was meant to pull accurate product listings for ad recommendations, "not about scraping news sites." On October 17, Meta deleted the clause outright and had spokesperson Daniel Roberts tell The Verge that "the crawling terms we consolidated only applied to our ads and business products" and that the change "does not change how Meta uses web crawlers." TechCrunch confirmed the removal the same day.

The removal changes the contract, not the crawling

Meta's own framing is that nothing about its actual web-crawling behavior changed, only the terms page. That leaves the underlying question the clause raised, whether Meta's commercial terms can override a site's robots.txt signal, unresolved rather than settled. Meta told The Verge that "content owners can use robots.txt to signal their preferences" and that the update doesn't "give us new rights that supersede those preferences," but that assurance now rests on a statement rather than contract language a publisher could point to. Operators who manage AI-crawler access through robots.txt or similar signals gained no new enforcement mechanism here; the practical lesson is that a platform's terms-of-use page, not just its public statements, is the document worth re-checking after a reported dispute like this one.

Context and reactions

The episode sits alongside Meta's existing paid licensing deals with CNN, Fox News, USA Today, and People Inc. for AI training content, deals the now-deleted clause would have let Meta bypass for any other business-linked site, per Nieman Lab's reporting as relayed by The Verge. It also lands a month after Meta launched its Meta One subscription, which charges businesses and creators to include links in posts and reels, a contrast Nieman Lab drew explicitly: Meta would have been charging for the same links it could crawl for free. Meta has prevailed in at least one major US copyright suit over AI training data, while other publisher suits against AI companies, Meta included, remain active.

What to watch next

Whether any publisher or regulator treats the quick reversal as tacit confirmation the original clause carried real legal risk, and whether Meta's commercial terms get another look once the October 30 effective date for the rest of the revision arrives.

Sources