Social Rocket
Update from 26 October 2026: Resistance to Meta's planned website crawling clauses is also being organised through industry associations. The US publishers' association News/Media Alliance called on Meta in an open letter dated 22 October to withdraw the controversial terms; European publishing organisations and media houses have also raised objections. The associations argue that a Meta business account should not automatically count as consent to the comprehensive scraping of editorial websites. No regulatory prohibition can be inferred from this. Meta's postponed start date remains, according to current confirmation, 20 November 2026. Further background: Digiday and Reuters.
Update from 25 October 2026: On 23 October, Meta confirmed that the revised Commercial Terms and Self-Serve Ad Terms will now take effect on 20 November 2026 instead of 30 October. The delay follows complaints from publishers and advertisers to competition authorities in the EU and the UK. Meta denies that it is thereby granting itself new blanket AI training rights and intends to explain the rules to those affected. What has been postponed is the start date, not the debate over the crawler clause. Sources: Reuters from 23 October and Reuters from 22 October.
Meta has softened its new business terms after protests from publishers and lawyers – but the right to automatically read linked websites remains included. The revised Commercial Terms are set to apply from 20 November 2026. They affect not only large advertisers but also companies that use Meta products for business, for example via a Facebook page, an Instagram business profile, product catalogues or Ads Manager. For German marketing teams, the key question is therefore less the removed word "train" than what Meta may technically retrieve from its own domain.
What was changed – and what was not
In an earlier version, Meta was allowed, according to documentation and a comparison by Nieman Journalism Lab, to also use content from websites linked to business accounts to train its products. After criticism, Meta removed this explicit wording and limited the described purpose to advertising and business products as well as security functions. Automated crawling, indexing, analysing and recurring retrieval of web pages connected to the products or services offered remain provided for.
Meta told The Verge that the passage does not change its existing crawler practice. Content providers can continue to express their preferences via `robots.txt`; the change is not intended to allow news sites to be copied. Critics nevertheless consider the purpose "develop business products" too broad and warn that the clause itself does not expressly protect `robots.txt`. That is not a legally conclusive assessment.
Why smaller brands should also check their website
Anyone connecting their domain to a Facebook page, Instagram ads or a product catalogue should clarify the scope of the data made available with IT and the legal department. For online shops, publicly accessible product information can indeed be useful: current prices and availability help with ads and recommendations. Licensed images, paid specialist content or content behind access barriers must be assessed differently. The new clause does not mean that Meta can automatically access non-public customer data.
Practical check before 20 November:
1. Compare the current preview of the Meta Commercial Terms with existing contract and rights approvals; the wording has already been changed once after the fact. 2. Document in the web team which Meta crawlers should access which areas; check `robots.txt` and server logs technically. Meta describes the file as a preference signal, not a legally guaranteed access block. 3. Keep the difference between social previews, ad product data and AI access in mind: blanket blocking can also disrupt legitimate link previews or commerce processes. 4. For protected content, assess licence rights and actual access restrictions separately. A contract clause does not replace an individually reviewed legal basis.
What the change shows for paid social and publishers
Meta is tying advertising, product information and AI systems ever more closely together. How heavily these systems rely on external signals anyway is shown by our article on Meta's agentic Ads Manager. The question of data origin equally affects personalised Meta Muse agents and the measurement of content in AI systems.
Classification: Meta has responded to public pressure but has not abandoned data sourcing via business-linked websites. For companies, a documented decision on crawlers, rights and marketing dependencies now makes sense – not a hasty assumption that the issue is settled after the wording change.