Meta’s proposed business terms set off alarms among online publishers after language surfaced that appeared to give the company sweeping rights over websites connected to its business accounts.

The clause was due to take effect on October 30 and applied to businesses promoting content through Facebook, Instagram, WhatsApp and Messenger. It said Meta could crawl websites “associated” with those accounts and use outside material to “develop, improve, research,” and “train” its AI products.¹

For publishers, the concern was not merely that Meta might collect publicly available pages. The wording appeared to make access a condition of doing business on Meta’s platforms: companies that declined the new policies could lose access to Meta products for commercial purposes. That raised the prospect that news organisations could be drawn into supplying training material without a separate deal, even as Meta had pursued licensing agreements with outlets including CNN, Fox News, USA Today and People Inc.¹

The apparent breadth of the clause was the flashpoint. If a publisher’s site was linked to a business account, critics feared Meta could use its journalism for AI development while bypassing the negotiations and payments increasingly associated with AI-content licensing.¹

Then the language vanished. Meta removed the wording that may have allowed it to scrape news sites, according to reporting on the change.² The removal addressed the immediate clause, but it also underscored the pressure around platform terms: publishers are increasingly alert to contractual language that could turn their web presence into free fuel for generative AI.