A paragraph that was not there before
Meta’s updated Commercial Terms take effect on 30 October, and they add a paragraph the current version does not contain. The clause sits in the section on licences and rights granted to Meta, and it authorises the company, or any entity acting on its behalf, to use automated software to access, index, cache, analyse and crawl the domains and pages associated with a business’s products or services, and to pull updates from them periodically.
A second sentence covers what Meta may do with what it collects. It lists providing, developing, improving, researching, training and maintaining the integrity of Meta’s products. The word that matters there is “train”.
The permission is not limited to advertisers. Search Engine Journal reported that the October version reaches businesses that never buy an ad: business use is defined to include running a Page, sending marketing messages, maintaining a product catalogue, and using Ads Manager or Meta Business Suite.
What the lawyers read into it
Nieman Lab put the text to three legal experts, who said the wording is broad enough to cover news organisations that use Meta’s products commercially. Cornell law professor James Grimmelmann told the outlet the language amounts to permission for Meta to train AI on any content from those sites.
Two details drew particular attention. The clause covers Meta “or any entity acting on our behalf”, so it is not confined to Meta’s own crawlers. And it permits the creation of test credentials to reach the pages, which one expert told Nieman Lab would technically let Meta register as a user to get past a site’s registration wall. The terms do not define what a test credential is or how it may be used.
The terms also leave no partial exit. Nieman Lab quoted the update telling businesses that do not agree to stop using Meta’s products for commercial purposes. There is no checkbox that declines the crawling paragraph while keeping the Page.
Meta says the reading is wrong
Meta spokesperson Andy Stone responded on X on 16 October, calling a summary of the Nieman Lab story false and saying there is no change to how Meta uses web crawlers. He said the update exists so Meta can pull accurate information about products sold on third-party sites when it makes advertising and content recommendations, and that it is not about scraping news sites. He pointed publishers to robots.txt, which he said remains the way to signal a preference.
A second Meta spokesperson, Daniel Roberts, told Nieman Lab the update consolidates existing commercial and advertising terms and will not change how the company crawls publishers in future.
Where the two accounts actually diverge
Stone’s answer is about practice; the lawyers’ objection is about permission. Nothing in Meta’s statements denies that the clause grants the right its critics describe — only that the company intends to use it that way. A term of service is enforced on its text, not on a spokesperson’s post, and Nieman Lab reported that Meta declined to answer its questions about whether the terms apply to news publishers.
The robots.txt point has the same gap. Robots.txt is a request a crawler may honour, and the clause itself does not mention it. Grimmelmann’s concern, as Nieman Lab reported it, is that a contractual permission could be argued to sit above a blocking instruction rather than beneath it.
What to watch
The date is the thing. On 30 October the terms apply to every business account that has not stopped using Meta’s commercial products, which is close to all of them. Watch whether Meta revises the text before then — it had not as of 17 October — and whether any publisher trade body tests the clause, in Europe especially, where the AI Act’s transparency duties on training data have applied since 2 August.
