Meta’s business terms still leave AI scraping room, legal experts say

Meta removed the word “train” from new business terms, but legal experts say broad “develop” language could still cover AI training. Content operators should review their exposure before October 30th.

What happened

Meta has tweaked its business terms to remove language suggesting that it could freely train its AI models on news sites, such as The Verge. The revised terms still concern legal experts. The terms, which go into effect October 30th, originally stated that Meta “may use any content” from websites associated with a business account to “develop, research” and “train” Meta’s products. Legal experts told Nieman Lab that this could allow Meta “to argue its commercial terms supersede other blocking instructions,” such as robots.txt.

Following fierce backlash from the publishing industry, Meta removed the word “train” from that section. However, James Grimmelmann, professor of digital and information law at Cornell Law School, told Nieman Lab: “This does nothing to restrict the ‘associated with your products or services’ [language] that would let Meta scrape entire sites.” Meta spokesperson Daniel Roberts said content owners can use robots.txt and the update “is not about scraping news sites, and these updates don’t give us new rights that supersede those preferences.”

Key facts

• Meta’s new business terms go into effect October 30th; the initial wording said Meta “may use any content” from websites connected to a business account to “develop, research” and “train” its products.
• After publishing backlash, Meta removed the word “train,” but the revised text still allows Meta to use scraped content to develop any “business product.”
• James Grimmelmann said the change does not restrict “associated with your products or services” language that would let Meta scrape entire sites.
• Matt Topic, media and intellectual property attorney at Loevy & Loevy, said “develop” is broad enough to include “train,” and Meta is presumably developing AI business products.
• Meta’s statement says robots.txt remains a signal and the update gives no new rights that supersede those preferences.

Our analysis

The dispute matters because it turns on how broadly a platform can claim rights over content that site owners never directly uploaded to Meta. Meta says the terms are not about scraping news sites and that robots.txt preferences still matter. But the legal reading from Grimmelmann and Topic suggests that removing the word “train” is not the same as closing the door. A phrase like “develop any business product” can function as a catch-all, especially when the company is already building AI products.

For content operators, the uncertainty itself is a cost. Publishers and brand sites commonly use robots.txt to signal what scrapers may crawl; historically that file has been treated as a polite instruction, not always a legal barrier. If Meta’s business terms can be argued to override those instructions, the practical value of robots.txt is weaker for anyone who runs both a website and a Meta business account. This likely pushes content teams to treat terms updates as legal review triggers, not as routine acceptance.

The change is different from a clear rollback. Meta removed explicit “train” language under public pressure, but left broad development language intact. That pattern suggests the company is adjusting wording to reduce publishing industry backlash while retaining flexibility for future AI use. Operators should read the update as an unresolved rights question, not as a resolved safeguard.

What it means for operators

• Before October 30th, review Meta’s business terms and compare them against your own site terms, especially if your brand or creator business runs a linked website, portfolio, or store.
• Check robots.txt and other crawl controls, but do not treat them as a complete defense; Meta’s public statement says it does not supersede those preferences, while legal experts identify broader “develop” wording.
• For original editorial, client work, or exclusive content, get a legal review of the revised terms and document what you are willing to allow for AI development use.
• Monitor how Meta applies the terms after October 30th and what publishing industry groups do next; update internal content-sharing and syndication guidance accordingly.