Meta delays contested content terms: what publishers and AI businesses should do now

What changed on 23 October

The immediate change is timing. The revised terms are now due to take effect on 20 November 2026. Meta said it would provide further explanation of its crawling practices and make clear that content owners may use robots.txt to communicate preferences.

The European Commission welcomed the delay, according to Reuters. The European Publishers Council nevertheless maintained that operating a business page or advertising a subscription should not confer general rights over a publisher’s website or archive.

The postponement does not resolve the disputed interpretation. It provides a short period in which affected businesses can identify connected domains, preserve the applicable versions of the terms and decide whether their technical and contractual controls express the rights position they intend.

A complaint is not a competition-law finding

The Movement for an Open Web says that it has asked the Commission and CMA to intervene. Its allegation is that businesses may depend on Meta’s services and could be required to accept content-use conditions capable of strengthening products that compete with them.

Article 102 TFEU and section 18 of the Competition Act 1998 prohibit abuse of a dominant position and identify unfair trading conditions as a possible form of abuse. Those provisions do not make every broad platform term unlawful. An authority would need to assess matters including the relevant market, dominance, the conduct, its effects and any objective justification.

No published decision has made those findings here. Businesses should therefore describe the competition issue as a complaint under assessment, not as an established abuse by Meta.

• Article 102 TFEU
• Competition Act 1998, section 18

Contract permission and copyright exceptions are different questions

A platform may rely on contract wording as a source of permission. Separately, copyright legislation may permit particular acts without permission. The scope and conditions of those routes must not be collapsed into one another.

In the EU, Article 4 of the Digital Single Market Copyright Directive requires a text-and-data-mining exception for lawfully accessible material, but the exception is conditional on rights not having been expressly reserved in an appropriate manner. For online content, the Directive identifies machine-readable means as an example.

In the UK, section 29A of the Copyright, Designs and Patents Act 1988 is narrower. It addresses copies for computational analysis by a person with lawful access for non-commercial research. It is not a general commercial AI-training exception.

Robots.txt can therefore be relevant evidence of a machine-readable preference or reservation, but it is not a complete legal solution in every jurisdiction or for every use. Businesses should align technical instructions, contractual notices and licensing positions rather than assume that one file settles all questions.

• EU Digital Single Market Copyright Directive, Article 4
• Copyright, Designs and Patents Act 1988, section 29A

Why AI businesses should examine their own terms

The dispute is not relevant only to Meta or news publishers. AI businesses frequently connect product accounts, customer websites, repositories and third-party tools. Standard terms may purport to authorise collection or reuse of information beyond what a customer expects from the immediate service.

A clause can create legal, commercial and procurement risk even before a regulator or court determines its validity. Enterprise customers may object to unclear rights, security teams may block integrations and marketing claims about customer control may conflict with the contractual text.

Product, legal and technical teams should be able to explain which crawler or integration collects which material, for what purpose, under which permission, and how a customer can refuse or withdraw that use without losing unrelated functionality.

Practical actions before 20 November

1. Identify every Meta business product, advertising account, Page and integration used by the organisation.
2. Record the domains, applications and content repositories associated with those accounts.
3. Preserve the current and proposed terms, related notices and internal approval records.
4. Compare contractual permissions with website notices, licences, robots.txt rules and other machine-readable controls.
5. Separate Meta’s different crawlers and purposes rather than relying on a single allow or block decision.
6. Check whether blocking or limiting a crawler affects advertising, discovery, analytics or other necessary functions.
7. Review contracts with contributors, agencies, licensors and customers to confirm who may authorise AI-related use.
8. Document any objection or reservation clearly and through the channels available under the relevant terms.
9. Avoid stating that a regulator has prohibited the terms or found them unlawful unless an official decision does so.
10. For the business’s own AI services, narrow content-use clauses to identified products and purposes, and provide intelligible controls.

The compliance conclusion

The three-week delay is not a final outcome. It is a warning that permissions embedded in platform terms can become a competition, copyright and customer-governance issue at the same time.

Businesses should use the additional period to establish what they have authorised, what their technical settings communicate and whether their own customer terms meet the standard of clarity they expect from larger platforms.