Meta commercial terms taking effect on 30 October 2026 authorise the company to crawl websites linked to business accounts and use what it collects to train its AI products. The terms cover Facebook, Instagram, WhatsApp and Messenger. Businesses that decline can lose business use of all four.
What the New Clause Says
The change adds explicit crawling rights to the contract every business account accepts. Nieman Lab, which first reported the revision on 15 October 2026, quoted the new language as authorising Meta “(or any entity acting on our behalf) to use automated software (such as web scrapers, bots, robots, spiders, crawlers and user agents)” and to “create test credentials to access, index, cache, analyse, collect, retrieve or crawl the domains” tied to a business.
A second passage covers what Meta may then do with the material. It says Meta may use any content from those domains, web pages or URLs “to provide, develop, improve, research, train and maintain the integrity of, the Meta Company Products, services and features”.
Two things are worth separating. The crawling itself is not new; Meta has run web crawlers for years and documents them publicly. What is new is a contractual permission, granted by the site’s own operator, covering training as a named purpose.
Who Is Bound, and What Declining Costs
The terms apply to business and commercial accounts across Meta’s four main apps. There is no opt-out tick box and no transition period: the revision takes effect on 30 October for everyone still using those accounts.
Refusing has a stated consequence. The terms say businesses that do not accept them “may stop all access or use of our Products for business or commercial purposes” — in practice, giving up Meta’s advertising and business tools altogether. Benjamin Sobel, a law professor at the University of Wisconsin, put the position to Nieman Lab plainly: “If you want to use the Meta products that require acceptance of the commercial terms, you have to also acquiesce.”
The timing has drawn attention because Meta began charging for a related function weeks earlier. Meta One for Business launched on 15 September 2026, and some Facebook Pages have since been limited to two organic posts containing external links each month unless they subscribe, a restriction first tested in December 2025. Publisher Pages are exempt from that test, according to Social Media Today.
What Meta Says It Is For
Meta rejects the reading that this is about news. Andy Stone, the company’s communications chief, said “there’s no change to how Meta uses web crawlers” and that the update is “designed to make sure we’re able to pull accurate information about products for sale” on third-party websites.
Meta also told Nieman Lab it has no plans to use the terms as a legal basis for ignoring instructions that block crawling on news sites. That is a statement of intent, not a contractual limit, and the terms themselves do not mention such instructions.
What the Lawyers Told Nieman Lab
Three legal academics and practitioners reviewed the wording for Nieman Lab. None of them described a settled legal position; all three read the clause as broad.
- James Grimmelmann, Cornell Law School: “For news publishers, this clearly represents permission to Meta to train AI on any content from their sites.”
- Matt Topic, Loevy & Loevy: “It’s not clear whether a court would enforce this, but publishers should assume that it would be enforceable.”
- Benjamin Sobel, University of Wisconsin: Accepting the products means accepting the terms, with no middle option.
Trade bodies have objected on both sides of the Atlantic. The News/Media Alliance, which represents United States news publishers, called the policy “outrageous and probably illegal”. The European Publishers Council has urged the European Commission to examine the choice it says European publishers now face between allowing AI scraping and losing platform access. Neither body’s objection has produced a regulatory action so far.
How to Block Meta’s Crawlers
Blocking is done in a site’s robots.txt file, and it has to be done per crawler. Meta documents five, each with its own user-agent token, and a rule written for one sets no policy for the others.
| Crawler token | Meta’s stated purpose | Robots.txt note |
|---|---|---|
| meta-externalagent | Training foundation AI models, or indexing content directly | No bypass stated |
| meta-webindexer | Improving Meta AI search result quality | Allowing it lets Meta cite and link your content |
| meta-externalads | Improving advertising and business products | No bypass stated |
| meta-externalfetcher | Fetching individual links at a user’s request | May bypass robots.txt |
| facebookexternalhit | Crawling pages shared on Meta’s apps, for titles and previews | Might bypass for security or integrity checks |
To stop the training crawler specifically:
- Open or create robots.txt: It must sit at the root of the domain.
- Add a rule for the training crawler: A User-agent line naming meta-externalagent, followed by Disallow: /
- Repeat for any other token you want stopped: Add a separate block per crawler, since the tokens are matched individually.
- Leave facebookexternalhit allowed if you want link previews: Blocking it removes the title, description and thumbnail when someone shares your page.
- Wait up to 24 hours: Meta’s documentation states its crawlers may cache robots.txt for that long, so a change is not immediate.
Meta’s own crawler documentation states that Meta-ExternalFetcher “may bypass robots.txt because it performs fetches that were requested by the user”, and that FacebookExternalHit “might bypass robots.txt when performing security or integrity checks”. No bypass is documented for the training crawler.
Why Blocking May Not Settle It
The lawyers’ concern is not that robots.txt stops working. It is that a contract signed by the site operator could be argued to sit above it. Robots.txt is a convention that crawlers choose to honour, not a legal instrument, and the revised terms grant permission directly from the site’s owner.
Meta has said it does not plan to make that argument. Whether it could is untested, and no court has ruled on a clause of this kind. For a business that keeps its Meta accounts, blocking the crawlers and granting contractual permission to crawl are not mutually exclusive positions — they simply have never been tried against each other.
A related limit applies regardless. Material already collected is unaffected by a rule added now, and content syndicated elsewhere can be reached through other sources. The wider question of whether training on published work is lawful remains open, as our report on the AI training fair use ruling set out.
What Is Still Untested
- Enforceability: No court has considered whether this clause would be upheld.
- Scope in practice: Meta says product data; the wording says any content from the domain.
- Regulatory response: Publisher bodies have complained, but no regulator has opened a case.
- Interaction with blocking: Whether Meta would treat the contract as overriding a crawler block, having said it has no plans to.
- What counts as a linked domain: The terms refer to domains tied to a business’s products and services, a boundary Meta has not defined publicly.
As of 17 October 2026, the terms are published and dated, no legal challenge has been filed, and Meta has not amended the wording. Businesses reviewing what Meta’s systems can reach may also want our guide to what the Meta Muse AI agent can access and how to revoke it.
Meta Commercial Terms FAQ
When Do Meta’s New Commercial Terms Take Effect?
On 30 October 2026, for business and commercial accounts across Facebook, Instagram, WhatsApp and Messenger. There is no transition period and no opt-out setting.
Can a Business Opt Out?
Not while continuing business use. The terms state that businesses which do not accept them may stop all access to Meta’s products for business or commercial purposes, so declining means giving up the business tools.
Which Crawler Trains Meta’s AI?
Meta’s documentation names meta-externalagent as the crawler used for purposes such as training foundation AI models. Four other tokens cover search indexing, advertising, user-requested fetches and link previews.
Does Blocking in Robots.txt Still Work?
Meta’s documentation says to block a crawler by adding a disallow rule for its token, and lists no bypass for the training crawler. Two other crawlers, Meta-ExternalFetcher and FacebookExternalHit, are documented as possibly bypassing it.
What Does Meta Say the Change Is For?
Andy Stone said there is no change to how Meta uses web crawlers and that the update is designed to let it pull accurate information about products for sale on third-party sites. Meta also said it has no plans to use the terms to ignore crawler blocks on news sites.
Has Anyone Challenged the Terms?
No legal challenge had been filed as of 17 October 2026. The News/Media Alliance and the European Publishers Council have both objected publicly, and the European body has asked the European Commission to look at it.
