Meta's updated terms let it scrape news sites linked to business accounts

By Splitfeed EditorialLast updated Oct 16, 2026, 14:10 UTC

On or around October 16th, 2025, Nieman Lab reported that Meta had updated its terms of service to permit the company to scrape websites that are "associated" with businesses using its platforms. The updated terms, set to take effect on October 30th, explicitly state that Meta can use content gathered from those external sites to "develop, improve, research," and "train" its AI products. The policy spans all four of Meta's major platforms: Facebook, Instagram, WhatsApp, and Messenger. News publishers that maintain business accounts on any of those platforms are among those potentially affected.

The significance of the change lies in its reach and its mechanism. Rather than asking businesses for separate consent to harvest their external web content, Meta is folding that permission into the standard terms of service that any business account holder must accept to use the platforms. For news organizations, which have increasingly pushed back against AI companies using their journalism for model training without compensation or agreement, this creates a difficult position: maintaining a social media presence on Meta's platforms may now implicitly license their site content for AI development. The core dispute is whether burying this permission inside platform terms constitutes meaningful consent, or whether it amounts to a unilateral land-grab dressed in legal boilerplate.

None of the three camps, Pro-AI, Anti-AI, and Middle Ground, have published direct reactions to this story yet. Pro-AI voices would typically argue that scraping publicly accessible web content is standard industry practice, that Meta's terms are at least transparent about the intention, and that the data use ultimately improves products that benefit users. Anti-AI voices would be expected to frame this as another example of AI companies extracting value from publishers without fair compensation, compounding existing concerns about copyright and the sustainability of professional journalism. The Middle Ground camp would likely call for clearer opt-out mechanisms and argue that platform dependency should not function as implied consent for AI training.

The date to watch most immediately is October 30th, 2025, when the new terms formally take effect. Any legal challenge, regulatory inquiry, or organized response from news industry groups before or shortly after that date would clarify whether the policy survives in its current form or faces meaningful pushback.