Meta AI Training Clause Removed: What Businesses Must Know
Meta cut the word train from its proposed Commercial Terms after legal experts warned the wording could permit scraping business websites for AI training. The crawling permission remains, so businesses should audit their associated domains and robots.txt settings before the terms take effect on October 30, 2026.
The Short Answer
On October 17, 2026, Meta confirmed that the Meta AI training clause removed from its proposed Commercial Terms was real, and that the change had already been made. The deleted sentence had said Meta could use content from domains, web pages or URLs associated with a business account to provide, develop, improve, research, train and…
On October 17, 2026, Meta confirmed that the Meta AI training clause removed from its proposed Commercial Terms was real, and that the change had already been made. The deleted sentence had said Meta could use content from domains, web pages or URLs associated with a business account to provide, develop, improve, research, train and maintain its products. After legal experts warned that the wording could be read as permission to scrape publisher websites for AI training, Meta cut the word train and narrowed the stated purpose to advertising and business products. The revised terms still take effect on October 30, 2026, and the crawling permission itself remains in place. For any business with a Meta business account, that combination is the whole story: one alarming sentence is gone, but the permission to crawl associated websites is still there, and there is no separate opt-out.
Meta AI training clause removed: what actually changed
The dates matter because they show how quickly contract language can shift once experts read it closely.
| Date | What happened |
| October 15, 2026 | Nieman Lab reported the terms after a reader tip, noting that legal experts read the wording as permission to scrape sites tied to a business account. |
| October 16, 2026 | An archived preview still contained the sentence allowing content from associated domains to be used to train Meta products. |
| Late October 16 to October 17, 2026 | Meta removed the training sentence. The Verge and TechCrunch reported the removal on October 17, 2026. |
| October 30, 2026 | The revised Commercial Terms take effect. Businesses accept them by continuing to use Meta's commercial products. |
Meta spokesperson Daniel Roberts told The Verge that the crawling terms Meta consolidated only applied to its ads and business products, and that the change does not change how Meta uses web crawlers. Meta communications vice president Andy Stone also said on X that the update was meant to pull accurate product listings for ad recommendations.
What the original draft said, and why lawyers worried
The original draft gave Meta permission to use automated tools, described as scrapers, bots, spiders, crawlers and user agents, to access, index, cache, analyse and collect material from domains, web pages and URLs associated with a business's products or services, and to retrieve updates periodically. The separate sentence that caused the alarm then said Meta could use that content to provide, develop, improve, research, train and maintain Meta Company Products, services and features.
Cornell digital law professor James Grimmelmann told Nieman Lab that the terms clearly represented permission for Meta to train AI on any content from business sites, and possibly to deploy some of that content in competition with the publishers themselves. Legal experts also warned that the wording could be read as overriding a site's robots.txt file, the standard file websites use to signal which crawlers may access which pages. If a business had already blocked AI crawlers in robots.txt, a contract clause granting Meta access to its associated domains could arguably have overridden that signal.
The worry reached beyond publishers for a simple reason: almost every news organisation holds at least one Meta business account, and so does almost every business that advertises on Facebook or Instagram. The clause was not limited to media companies. Any shop, agency, creator or club using a Page for business purposes and owning a website fell inside it.
Meta AI training clause removed, but crawling permission remains
Here is the part that deserves the most attention. The October 17 version removed the separate sentence that expressly listed training, but it kept the crawling authorisation. The revised passage still permits Meta, and companies acting for it, to use scrapers, bots, spiders, crawlers and user agents to access, index, cache, analyse and collect material from domains, web pages and URLs associated with a business's products or services, and to pull updates periodically.
What changed is the stated purpose. Instead of a broad list ending in training, the revised text describes purposes connected to advertising and business products, alongside maintaining the integrity and security of Meta Company Products. The News/Media Alliance said the edit limits content use but not collection. That distinction is the key to reading the revised terms honestly: Meta no longer claims an express right to train on the crawled content in this clause, but it still claims the right to collect it.
It is also worth noting what the revised terms do not contain. According to Search Engine Journal, neither the updated Commercial Terms nor the self-serve advertising terms mention robots.txt. Meta has said separately that content owners can use robots.txt to signal their preferences, and that the update does not give Meta new rights over website content. When a contract and a press statement disagree, the contract is the document that will be quoted later.
What Meta says has not changed
Meta's consistent message is that nothing about its actual crawling behaviour changed. The company says the revision clarified that consolidated crawling provisions apply only to ads and business products, and that it does not change how Meta uses web crawlers. It also says site owners can still use robots.txt to signal their preferences.
First, Meta operates several distinct crawlers, and they do not all behave the same way. Meta's crawler documentation names Meta-ExternalAgent, which crawls for uses including training foundation AI models or indexing content, Meta-ExternalFetcher, which fetches links at a user's request, and FacebookExternalHit. Meta's documentation also states that Meta-ExternalFetcher may bypass robots.txt when retrieving a page at a person's request, and that FacebookExternalHit may do so during security checks.
Second, the wider legal background has not changed either. Meta prevailed in a high-profile copyright lawsuit over AI training data last year, while other cases involving AI companies remain active. Meta also holds paid licensing deals with publishers such as CNN, Fox News, USA Today and People Inc. for AI training content.
One deleted word changes the argument a company can make later
This is the original insight at the heart of the story, and it applies far beyond Meta. Deleting one word from a contract changes the argument a company can make later, even if nothing changes in its crawlers today. Before the edit, Meta could have pointed to an express sentence and argued that businesses using its commercial products had agreed to training use of content from their associated domains. After the edit, that sentence is gone, and any future claim would have to be built from broader language about advertising and business products.
That is why the removal matters even though Meta insists its behaviour is unchanged. Contracts are not descriptions of what a company does today. They are the set of arguments a company is allowed to make tomorrow, in a negotiation, in a dispute, or in court. A clause that is never enforced still shapes what the other side believes it has agreed to, and a clause that is deleted removes that leverage.
The reverse is equally true. The permission that remains is the part to audit. Crawling authorisation, periodic retrieval of updates, and access to domains associated with a business account are still in the text, and they govern what Meta's systems actually collect from your website.
For businesses that run paid campaigns, this connects directly to how advertising platforms read and use website content. If your product pages, landing pages and blog posts feed ad recommendations and business tools, then crawler access to those pages is part of your marketing infrastructure. Reviewing it alongside your broader digital marketing and SEO strategy helps you decide which pages should be easily readable by platforms, and which should be restricted.
Practical checks for your business before October 30
1. List every domain associated with your Meta business account
Start in Business Suite and your advertising account settings. Note every domain you have verified, every website linked from your Page, and every product catalogue feed that points at your site.
2. Review your robots.txt file
Open your robots.txt file and check which Meta user agents are named, if any. Meta says content owners can use robots.txt to signal their preferences, so make sure your file actually states them. If you want to block AI training crawlers, name Meta-ExternalAgent explicitly, since that is the crawler Meta documents as crawling for training foundation models. Keep in mind the documented exceptions: fetchers acting on a user's request and security checks may bypass robots.txt, so do not treat robots.txt as an absolute barrier.
3. Audit what lives on those associated domains
Walk through the pages a crawler could collect: product descriptions, blog posts, case studies, team pages, downloadable resources and any staging or test content that is publicly reachable. Ask whether any of it is material you would not want used to improve advertising products, such as unreleased product details or client work under confidentiality. If such content exists, remove it or restrict access before October 30. Businesses handling sensitive client or customer data should treat this as part of their wider cybersecurity posture.
4. Separate public marketing content from restricted content
Most businesses want their public marketing pages to be crawlable, because visibility drives discovery and ad relevance. The problem is rarely the public blog post. It is the page that was never meant to be public. Check for indexed staging subdomains and old campaign landing pages. If a page should not be collected, it should not be publicly reachable.
5. Decide what acceptance means for you
There is no separate consent step and no partial opt-out in the revised terms. Businesses that disagree are told, in effect, to stop using Meta products for business or commercial purposes, which for many means giving up Pages, advertising and Business Suite tools. Most businesses will reasonably choose to continue. If you continue, do so deliberately: record that you reviewed the revised terms and set a reminder to re-check them when Meta next updates them.
What this means for AI, privacy and platform power
The episode also illustrates a broader shift in how platform terms work. A business account now ties together advertising, messaging, product listings and website crawling, and one clause can reach across all of them. The same pattern appears wherever AI systems read content beyond the app they live in, which is why platform access should be treated as a data governance question, not only a marketing one. Our earlier piece on AI agents reading browser tabs and the privacy risks for businesses covers a related version of the same problem.
Data ownership is the other half of the story. When a platform collects content from your website, questions follow about who controls that copy and what it may be used for. Those questions are worth answering in your own data policy before a platform's terms answer them for you. Our guide on complete data privacy and ownership explains why clear ownership rules matter more than ever.
Finally, keep the episode in proportion. Meta removed the training sentence, Meta says its crawling behaviour is unchanged, and Meta says robots.txt still applies. None of that means Meta has stopped crawling, and none of it means robots.txt was overridden.
Frequently asked questions
Did Meta stop crawling business websites?
No. Meta removed the sentence that expressly mentioned training, but the revised terms still authorise Meta to use scrapers, bots, spiders, crawlers and user agents to access, index, cache, analyse and collect content from domains associated with a business account. Meta says the change does not alter how it uses web crawlers.
Does the removal mean Meta cannot train AI on my website content?
It means this particular clause no longer expressly grants training use. It does not settle the wider question of AI training, which is governed by other terms, licences and copyright law.
Can I opt out of the crawling permission while keeping my Meta business account?
No separate or partial opt-out is provided in the reported terms. Businesses that disagree are effectively told to stop using Meta products for business or commercial purposes. Your practical controls are robots.txt signals, crawler-specific rules, and limiting what content is publicly reachable on associated domains.
Does robots.txt still apply after the revision?
Meta says content owners can use robots.txt to signal their preferences, and that the update does not give Meta new rights over website content. However, the revised terms text itself does not mention robots.txt, and Meta's documentation notes that some fetchers may bypass robots.txt when acting on a user's request or during security checks.
When do the revised terms take effect, and how do I accept them?
The revised Commercial Terms take effect on October 30, 2026. There is no separate consent step. Businesses accept them by continuing to use Meta's commercial products after that date.
What should my business check first?
Start with the domains associated with your Meta business account, review your robots.txt file for Meta's crawlers, especially Meta-ExternalAgent, and audit associated domains for content that was never meant to be public.
If you would like help reviewing your website's crawler settings and data exposure, our team can audit your site and tighten what platforms can reach. Contact Skills Technology Solutions to arrange a review of your website and Meta business setup.
Related Articles
Keep reading — more guides you may find useful.
Related Services
Services connected to this guide.
