From 30 October 2026, Meta's new Commercial Terms allow the platform to crawl and copy websites linked to business pages, adverts and Business Suite accounts, in order to train its own artificial intelligence products. The change, reported on 16 and 17 October 2026 by info-lux.com, Entrevue and trade outlet PPC Land, is not aimed at any single country but affects every company running a Facebook page, an Instagram advert or a catalogue tied to Meta, including in Belgium. There is no opt-out: continuing to use these tools after 30 October counts as acceptance.
1. What Meta's new terms actually change
A clause added to Section 1 of Meta's Commercial Terms allows the company to use automated software and test credentials to access, index, copy and re-crawl websites linked to a business account's products or services. The stated purpose is to provide, develop, improve and train Meta's products, according to PPC Land, which quotes Meta spokesperson Andy Stone.
The rule applies to anyone managing a page, running an advert, publishing a product catalogue or using Meta Business Suite. There is no partial opt-out: staying silent counts as acceptance from 30 October, PPC Land reports.
2. Why this also concerns Belgian companies
The change is not announced for any specific country. It applies everywhere Meta operates, including to Belgian companies running a Facebook page, an Instagram advert or a WhatsApp Business account linked to their own website. The site in question is the one the business itself has linked to its account, not a third-party site chosen by Meta.
Jason Kint of Digital Content Next says Meta appears to be granting itself the right to scrape websites across the open web. Lawyer Matt Topic adds that a court could find the clause enforceable, even though its real scope remains uncertain, according to PPC Land on 17 October 2026.
3. What a company can still decide
Meta says it respects websites' robots.txt files, but this guarantee does not appear in the text of the new terms, according to Entrevue. Before 30 October, a Belgian company can still check which sites are actually linked to its Meta accounts and adjust the access rules on its own server.
Leaving Meta altogether is not realistic for most Belgian SMEs, which rely on this audience for advertising and visibility. The useful decision concerns what content sits on the linked website, not whether to abandon the platform.
The AI lens, humans first
Meta's crawler carries out the exploration and collection of content automatically, at scale, with no human involved at the moment it passes through. But deciding which sites to leave accessible, how to set the robots.txt file, and whether the business's Meta account should stay as it is after 30 October, is a human call, legal and strategic. AI executes. Expertise decides, and keeps watch.
This week
This week, list every website linked to your Meta accounts: Facebook page, ad account, product catalogue, Business Suite. Check your website's robots.txt file and decide whether to block or allow the Meta-ExternalAgent crawler. Decide before 30 October whether to keep your pages and adverts as they are or remove sensitive content from the linked site.
Vistalaro Build audits the robots.txt file and the websites linked to your Meta accounts to regain technical control before 30 October. Vistalaro Reach checks that your live Meta campaigns do not expose more content than necessary on the website linked to your adverts. Vistalaro Pilot assesses the GDPR and content ownership risks with you before deciding whether your presence on Meta needs to change.
Do you know which websites are currently linked to your Meta accounts?
An audit before 30 October beats finding out too late what they contain.
Let's talk- PPC Land, Meta grants itself the right to crawl advertisers' websites for AI (17 October 2026)
- Entrevue, Meta modifie ses conditions d'utilisation pour aspirer le contenu des sites professionnels sans licence (17 October 2026, in French)
- Info Lux, Page Facebook ou pub Instagram : dès le 30 octobre, les conditions de Meta autorisent le crawl de votre site pour entraîner ses produits, ce qu'une entreprise de la région peut vérifier (16 October 2026, in French)