Meta is rewriting its Commercial Terms so that anyone who uses its products for business, even without buying a single ad, hands it the right to send automated crawlers to their websites and train its products on what they find. The changes take effect Oct. 30. Search Engine Journal’s Matt G. Southern summarized them on Oct. 16, following an Oct. 15 report from Nieman Lab.
The trigger is ordinary activity. Running ads counts, but so does managing a Facebook Page or using Meta’s business tools. Meta’s existing advertising contracts already grant crawl rights over advertisers’ sites, so the October version extends a familiar clause to a much larger group of account holders.
What the clause says
The new paragraph authorizes Meta, or anyone working on its behalf, to run automated software, including scrapers and bots, that gathers material from sites and pages connected to what a business sells or offers. Links that enter Meta’s systems because the business, or its customers, interacted with Meta products fall under the clause too. Meta may also “create test credentials” to reach those pages. Listed purposes include developing, improving, researching, and training Meta’s products.
An identical crawl grant sits in Meta’s updated ad-purchasing contract, the Self-Serve Ad Terms, on the same Oct. 30 date. Continuing to use Meta’s services for business after that date counts as acceptance. A business that objects can stop using them.
Two readings of the same language
Nieman Lab asked three legal experts to review the terms. They concluded the wording could support an argument that Meta may scrape news sites and use the material to train its AI products. Matt Topic, a media and intellectual property attorney who represents The Intercept in its copyright case against OpenAI, told Nieman Lab that no one can promise a court would enforce the provision, but publishers should assume it will be enforced. James Grimmelmann, a Cornell law professor, noted that test credentials might technically allow Meta to enroll as a bogus user and pass through registration barriers.
Meta rejects that framing. On Oct. 16, spokesperson Andy Stone answered on X, called the summary false, and wrote that the update exists so Meta can pull accurate details about products sold on outside sites when it recommends ads and content. In his words, it is “not about scraping news sites.” Meta also told Nieman Lab, and Stone repeated on X, that publisher crawling will work as before and that robots.txt remains available for managing access.
Two gaps keep the dispute open. Robots.txt is absent from both the Commercial Terms update and the ad-terms update. And Meta declined to tell Nieman Lab whether test credentials could get around paywalls or other paid content.
Robots.txt and Meta’s crawlers
The legal experts told Nieman Lab Meta could use the wording to argue that its business contract takes priority over other instructions to block crawlers. Whether a court would agree is untested. The terms also leave open whether a site’s robots.txt rules narrow the grant a business accepts when it agrees.
Meta’s crawler documentation describes separate agents. Meta-ExternalAgent crawls for purposes that include training foundation AI models. Meta-WebIndexer feeds Meta AI search, and Meta says permitting it makes citing and linking your pages in Meta AI answers easier. That split means a site can block one while permitting the other. The documentation also says Meta-ExternalFetcher, which retrieves a page on a user’s behalf, may ignore robots.txt, and FacebookExternalHit, which builds link previews, may do so during security or integrity checks.
The European Publishers Council is writing to the European Commission about the terms. Its worry is that Meta could assert rights over a news story each time someone posts its link on social media.
What site owners can do
Start with scope: any brand with a Facebook Page or a Business Suite account is covered, ad spend or not. Then pull your server logs for the user-agent strings Meta documents and note which pages each one requests. Check how your robots.txt treats Meta-ExternalAgent versus Meta-WebIndexer, because the answer decides whether you stay eligible for citations in Meta AI while declining training.
Operators have until Oct. 30 to decide whether business use of Meta is worth the new grant. Teams that cannot get a legal read by then should at least document their current robots.txt rules and crawl baselines, since those records will matter if a dispute over the clause’s reach ever arises.
Originally reported by Search Engine Journal (Matt G. Southern, October 16, 2026), citing Nieman Lab’s October 15, 2026 report and a post on X from Meta spokesperson Andy Stone.
