PolicyLaw, regulation & governance
Meta says it may delay its new business terms again, past the Nov. 20 date, while industry consultation continues
Meta's communications head says its new Commercial Terms, already pushed to November 20, could be delayed further while industry consultation continues. The draft would treat continued use of Meta's business tools as consent to crawl linked websites. Legal experts say the wording could still allow AI training.
From the The Verge · AI article of October 28, 2026 and cited sources: Andy Stone, Meta's communications head, said in a post on X that Meta won't change the policy until it has fully consulted with industry groups. [1] [4] [7]
SOURCES[1] Meta says it’ll probably delay its controversial new… theverge.com[4] PPC Land: Meta's new terms let it crawl business websites… ppc.land[7] WSAU (Reuters): Meta to delay new commercial terms… wsau.com
AT Pulse research deskPublished · Updated · 5 min read · AI-assisted analysisAI-assisted, not human-edited. Methodology Corrections Who owns this site
- What happened
- Andy Stone, Meta's communications head, said in a post on X that Meta won't change the policy until it has fully consulted with industry groups. [1] [4] [7]
- Original source
- The Verge · AI article, independent perspective
- Brief by
- AT Pulse research desk AI-assisted; publisher and contact on the desk page
- When
- October 28, 2026 article published
- Covered
- Oct 30 edition
- Evidence
- 9 sources across 7 sites
- 1Meta says it’ll probably delay its controversial new…theverge.com
- 2Nieman Lab: Legal experts warn Meta's new terms still…niemanlab.org
- 3Nieman Lab: Meta's new commercial terms will allow the…niemanlab.org
- 4PPC Land: Meta's new terms let it crawl business websites…ppc.land
- 5Media Copilot: Meta delays scraping terms to Nov. 20 amid…mediacopilot.ai
- 6Tech Insider: Meta Commercial Terms Delayed to Nov 20, 2026tech-insider.org
Reach: not measured yet
Andy Stone, Meta's communications head, said in a post on X that Meta won't change the policy until it has fully consulted with industry groups. He said Meta will delay again if that consultation runs past November 20. The current terms, in place since March 2024, stay in force until then. [1] [4] [7]
The first draft let crawlers regularly pull content from sites linked to a business account, to develop and train Meta's products. Meta later removed the word "train" and limited use to advertising and business products. Legal experts say the revised text still allows broad scraping. Meta says site owners can still block its crawlers with robots.txt, but that commitment is not written into the terms. [2] [3] [4] [5] [7]
Robots.txt offers limited control. The terms never mention it, and experts say a signed contract could be argued to override it. Meta-ExternalAgent also uses one robots.txt token for both AI training and indexing, so a site cannot allow one and block the other. [4]
Under the drafted terms, simply continuing to use Meta's business tools counts as consent to crawling, and robots.txt cannot separate training from indexing.
Read the full assessment
As a result, contract terms rather than technical blocking may decide whether publisher content ends up in AI systems.
Research is done; writing the dossier now.
Executive brief
Meta now says the new business terms it planned to apply to almost every news publisher on Facebook and Instagram may not take effect for an indefinite period. Meta's comms head, Andy Stone, said consultation with industry groups could run past the already-delayed November 20 date, and that the existing terms stay in force until then (The Verge). The draft terms would have treated continued use of Meta's business tools as consent to crawl the business's website, with no opt-out. Legal experts say that even after Meta's edits, the wording could still support AI training on publisher content (Nieman Lab).
What changed and event timeline
Legal experts flag a scraping clause
Three legal experts told Nieman Lab that the terms could let Meta scrape news sites and train AI on their content, with no opt-out (;).
Meta drops the word "train."
Meta removed the "train" wording and limited future use of scraped content to "advertising and business products." Reports give different dates for this change (;).
Complaints go to EU and UK regulators
Publishers and advertisers, including the Movement for an Open Web, complained to EU and UK competition authorities. They called the terms anti-competitive ().
Start date moved to Nov 20
Meta said the three extra weeks are for explaining "what's changing and what's not," including how its crawlers work. Media Copilot dates this announcement Oct 27 (;).
A second, open-ended delay
In a post on X, Stone said Meta won't change its policy without full industry consultation, and that it would delay again if consultation runs past Nov 20 ().
Capabilities and access
This is a policy change, not a model release. No model or version is named.
Read the full section
This is a policy change, not a model release. No model or version is named.
- Scope: Meta's Commercial Terms and Self-Serve Ad Terms. They cover every business using Facebook, Instagram or WhatsApp business tools. Continuing to use those tools after the effective date counts as acceptance (PPC Land).
- Crawlers: Meta-ExternalAgent uses one robots.txt token for both AI training and indexing. Meta-WebIndexer builds the index behind Meta AI answers. FacebookExternalHit fetches link previews (PPC Land).
Technical analysis for researchers and developers
- The original Section 1 authorized "web scrapers, bots, robots, spiders, crawlers and user agents" to access content linked to a business.
- Experts say a signed contract could be argued to override it (PPC Land).
- Because Meta-ExternalAgent uses one token for training and indexing, a site can't use robots.txt to allow one and block the other.
Read the full section
- Clause wording. The original Section 1 authorized "web scrapers, bots, robots, spiders, crawlers and user agents" to access content linked to a business. It let Meta "periodically pull updates" to "provide, develop, improve, research, train and maintain" its products (PPC Land).
- Contract vs. robots.txt. The terms never mention robots.txt. Experts say a signed contract could be argued to override it (PPC Land).
- Shared crawler token. Because Meta-ExternalAgent uses one token for training and indexing, a site can't use robots.txt to allow one and block the other.
- What's undocumented. No public material describes how scraped content would be filtered, logged or kept separate from training data.
Claims and evidence
- Meta's position: the crawling terms "apply only to ads and business products," and site owners can still use robots.txt (WSAU/Reuters).
- Experts' reading: the revised text still allows mass scraping (Nieman Lab).
- Wording: the terms reportedly didn't change during the first delay (Relevant Audience).
Read the full section
- Meta's position: the crawling terms "apply only to ads and business products," and site owners can still use robots.txt (WSAU/Reuters). These are company statements, and the terms themselves don't contain the robots.txt commitment.
- Experts' reading: the revised text still allows mass scraping (Nieman Lab).
- Wording: the terms reportedly didn't change during the first delay (Relevant Audience).
- Political pushback: Sen. Blackburn (R-TN) accused Meta of trying to "exploit" creators and businesses, and Sen. Schiff (D-CA) urged it to reverse the change (The Verge; Nieman Lab).
Context and prior work
- The previous terms took effect March 6, 2024 and had no crawling language (PPC Land).
- Groups that pushed back include the News/Media Alliance, WAN-IFRA, the UK's News Media Association and European publishers.
- The European Publishers Council says any use beyond running the requested service should need separate, express permission from the rights holder (WSAU/Reuters).
Read the full section
- The previous terms took effect March 6, 2024 and had no crawling language (PPC Land).
- Groups that pushed back include the News/Media Alliance, WAN-IFRA, the UK's News Media Association and European publishers. European publishers wrote a joint letter to the European Commission (Media Copilot).
- The European Publishers Council says any use beyond running the requested service should need separate, express permission from the rights holder (WSAU/Reuters).
Limitations, safety and contested findings
- Conflicting dates: reports differ on when "train" was removed (Oct 16 or Oct 17) and when the Nov 20 delay was announced (Oct 23 or Oct 27).
- Regulators: the European Commission welcomed the delay and said it would keep monitoring. No source reports formal enforcement action.
- Corroboration: no independent technical evidence, such as crawl logs, has been published showing Meta actually used publisher content to train models.
Read the full section
- Conflicting dates: reports differ on when "train" was removed (Oct 16 or Oct 17) and when the Nov 20 delay was announced (Oct 23 or Oct 27).
- Regulators: the European Commission welcomed the delay and said it would keep monitoring. The UK CMA said it would "continue to scrutinize" the terms (Media Copilot). No source reports formal enforcement action.
- Corroboration: no independent technical evidence, such as crawl logs, has been published showing Meta actually used publisher content to train models.
Business and practitioner implications
- Publishers: check which Meta business accounts you run. Under the drafted terms, simply continuing to use them would count as consent.
- Crawl controls: review robots.txt rules for Meta's crawlers. Contractual consent may still matter more than the technical signal.
- AI developers: getting training permission through bundled terms of service now draws regulatory and bipartisan political attention.
Read the full section
- Publishers: check which Meta business accounts you run. Under the drafted terms, simply continuing to use them would count as consent.
- Crawl controls: review robots.txt rules for Meta's crawlers. Contractual consent may still matter more than the technical signal.
- AI developers: getting training permission through bundled terms of service now draws regulatory and bipartisan political attention.
- Watch next: any revised wording that comes out of Meta's consultation.
The source trail.
Sources (9)
Meta says it’ll probably delay its controversial new business terms again
Article text retrieved; extracted text may omit tables or interactive elements.
theverge.com