Meta’s new terms may allow it to train AI on news sites for free
The Verge Emma Roth ● Covered by 2 sources
Meta’s new business terms may let it use publishers’ site content to train AI. If that holds up, news outlets could end up feeding Meta’s models for free.
Based on reporting by The Verge, Emma Roth — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Meta has quietly updated the terms for business accounts, and the change looks big. According to a report from Nieman Lab, the new language appears to let Meta crawl and scrape websites “associated” with businesses that use its platforms, including online publishers.
The timing matters too. The policy takes effect on October 30th, and it says Meta can use that outside material to “develop, improve, research,” and “train” its AI products. That covers a lot of ground, and it’s not hard to see why publishers are nervous.
The rule applies to business accounts that promote content across Facebook, Instagram, WhatsApp, and Messenger. Nieman Lab says the wording seems to give Meta a way around licensing deals, letting it train models on material posted by online publishers without paying for a separate agreement.
Meta hasn’t finished the public story here — the source text cuts off mid-sentence — but the direction is clear enough. This is another reminder that platform terms can do a lot of policy work before anyone has time to notice.
My take — AI-written commentary, not fact-checked reporting
This is the classic platform move: write the permission slip first, argue about fairness later. Publishers have spent years feeding social traffic machines; now the machine wants the archives too. If AI companies can turn terms of service into a free content buffet, they’ll do it every time.
Read more about this at: The Verge