Meta AI training terms update — Meta has indefinitely delayed a contentious update to its commercial terms that would have forced Facebook business page owners to permit AI training on external websites linked from their pages. The decision follows weeks of mounting opposition from publishers, legal experts, regulators, and lawmakers across multiple continents.
What the Original Proposal Entailed
The proposed changes to Meta’s AI training terms update would have required any business maintaining a Facebook page to allow Meta to crawl external sites linked from that page. Legal analysts warned that the vague language could let the company disregard robots.txt preferences and harvest enormous volumes of off-platform content for AI development. Because nearly every news publisher operates a Facebook page, critics cautioned that Meta could sidestep millions in licensing fees if the terms took effect as written.
Businesses had no means to opt out. To keep their pages, they would have had to consent to Meta crawling any linked external sites. The terms even obligated publishers to generate test credentials so Meta’s bots could reliably get past paywalls.
Mounting Global Scrutiny
The changes were first scheduled to begin on October 30. Meta later moved that date to November 20 as criticism intensified from users, courts, and regulators worldwide. Courts in Brazil and Norway issued preliminary injunctions blocking the terms after organizations in those countries filed lawsuits. The United Kingdom’s Competition and Markets Authority also launched a formal investigation into whether Meta was abusing a dominant position to secure valuable AI training data.
In the United States, bipartisan concern emerged. Sens. Adam Schiff (D-Calif.) and Marsha Blackburn (R-Tenn.) sent letters urging Mark Zuckerberg to withdraw the terms.
Industry Reaction and Meta's Response
Theo Bamber, chief executive of the UK’s News Media Association, voiced sharp criticism in a LinkedIn post. He described the company’s confidence that everyone would accept its actions as staggering, and he questioned whether Meta believed the public could not read the terms. Bamber argued that swapping the word “train” for “develop” offered little comfort to publishers, since the latter term carried the same threat. He noted that the language would let Meta use all the content it scrapes to develop “any business product.”
On October 28, Meta’s vice president of communications, Andy Stone, announced on X that the company would withdraw the terms, at least for now. He said Meta had heard publishers’ feedback and cared deeply about its partners, adding that it did not plan to make policy changes without full consultation with industry groups and leaders. Stone also indicated that because the consultation could extend beyond November 20, implementation of the updated terms would be delayed.
In a subsequent message to business page owners, Meta offered a reminder that its crawling practices remained unchanged and that content owners could continue to signal their crawling preferences using robots.txt. The company has consistently maintained that it would never use its terms of service to circumvent robots.txt.
Even after promising to pull the terms and consult with publishers, Meta still drew criticism. As regulatory and court scrutiny intensified, the company took steps to address publishers’ concerns but did not include any specific language about respecting robots.txt preferences in its revised terms of service.
What Comes Next
Meta’s existing terms remain in effect while the company engages with stakeholders. The indefinite pause leaves open the question of whether the company will eventually revive the update with clearer protections for publishers, or abandon the approach entirely. For now, the controversy has underscored the growing tension between AI development and the rights of content creators.
