Meta has backed away, for now, from a change to its commercial terms that critics said would let it use content from external websites to train AI. After weeks of backlash and a failed attempt to calm it, Meta has indefinitely delayed the update, according to a report by Ars Technica, headlined that Meta abandons a sketchy commercial terms update, and coverage by 404 Media and San Jose Inside. Ars Technica’s article could not be fetched here, so details come from these other reports.
What was proposed. The update would have required Meta Business account users to allow extensive AI training on external websites linked on Facebook, according to the reports. One report says the terms would have let Meta use any content from those domains, pages or URLs to provide, develop, improve, research, train and maintain the integrity of its products. Publishers feared that linking to their own sites would give Meta rights over their content.
Earlier changes. After pushing the update to November, Meta removed the word train from a list of allowed use cases and limited those uses to advertising and business products, according to a Nieman Lab report cited in other coverage. That change did not end the criticism, and the terms were then put on pause.
What Meta said. In a statement, Meta said its existing terms remain in effect while it consults stakeholders, and that it will share a further update when it has more to share, according to the reports. The statement does not say whether the AI training clause will return in a new form.
A conflicting account. One Kenyan news site says Meta withdrew the provision outright, which does not match the indefinite delay described by the other reports. The more consistent account is a pause, so the issue may not be closed.
Why publishers cared. News and media sites rely on Facebook links to reach readers, and many are already in disputes over how AI companies use their work. A requirement to accept AI training rights as a condition of using Meta’s business tools would have put them in a difficult position. The backlash shows how sensitive the question of consent for AI training has become.
Why it matters. The episode is another sign of the fight over who controls the web content that AI models learn from. It follows lawsuits such as the New York Times’ complaint against Brave, covered earlier on this site. For publishers and businesses in the US and UK, the practical step is to read platform terms carefully when they change, since defaults can shift how content may be used.
What to watch. Look for any new version of the terms, a consultation process, and whether other platforms try similar language.
Sources: 404 Media and San Jose Inside, with the Ars Technica report by headline. Details are as reported by those outlets and Meta’s statement.
Meta headquarters sign (file image), Wikimedia Commons, CC BY-SA 4.0
