Meta Drops the AI-Training Line From Its Business Terms — But the Crawling Stays
On October 17, 2026, Meta revised its pending Commercial Terms, removing a sentence critics said could let it use business-linked website content to train AI products. The terms still take effect October 30 — and the broad crawling permission behind the row remains.
If your business has a website plus a Facebook Page, Instagram business account, or Meta ads account, this fight matters to you. The fine print is where the real work is.
What happened
Nieman Lab surfaced Meta’s updated Commercial Terms on October 15. The original text authorized Meta, or anyone acting for it, to use scrapers, bots, and crawlers — and even test credentials — to crawl domains and URLs tied to a business account.
The alarming sentence came next: “Meta may use any content from such domains, web pages, or URLs to provide, develop, improve, research, train, and maintain the integrity of, the Meta Company Products, services, and features.”
Three legal experts told Nieman Lab that wording could be read as permission to scrape news sites for AI training. Attorney Matt Topic called it “a remarkable provision to bury in terms of service,” giving Meta material “that would likely cost many millions of dollars in a licensing market.”
Meta pushed back the same day. Spokesperson Andy Stone said nothing changed in how Meta uses crawlers, that robots.txt preferences still apply, and that the update aimed to pull accurate product info for ads — “not scraping news sites.”
Then the wording changed. Archived captures show the “train” sentence present October 16 and gone by October 17. The revised clause limits the purpose to advertising and business products, plus integrity and security. Meta: “We’ve heard you and want to clarify: the crawling terms we consolidated only applied to our ads and business products, and we’ve clarified that in the terms themselves. It does not change how Meta uses web crawlers.”
What did not change matters as much. Crawling permission remains, continued use counts as agreement, and there is no opt-out short of quitting Meta’s products. Some publishers say training remains “still potentially on the table” since the revision removes the word without ruling it out. Those readings are untested — read the final terms yourself before October 30.
Why it matters for small business
This was never only a publishers’ story. The terms cover businesses using Meta products commercially — ads, catalogs, Pages, Groups, Ads Manager, and Business Suite. Any shop or local business running a Page plus a website falls inside the crawling clause.
Two takeaways. First, your website content is part of the deal: the clause reaches beyond what you upload to Meta, to domains “associated with” your products or services. Second, the pushback worked — objections within days narrowed the text before it took effect. Pay attention during the notice window. Control is also technical: the lever Meta points to is robots.txt, telling automated visitors what they may fetch — much like knowing what automated tools acting on your accounts can reach. And as AI rules tighten worldwide, expect more terms fights.
Pricing, availability, and what to do next
No price change — this is a terms revision, not a product launch. The revised terms take effect October 30, 2026. Three checks before then:
1. Read the revised terms. Open the current Commercial Terms from Meta’s legal pages or Business Suite notices. Confirm the “train” sentence is gone in the version applying to you. Do not rely on summaries — including this one — for legal decisions.
2. Decide your robots.txt policy. Meta names three crawlers: meta-externalagent (training models or indexing); meta-externalads (ads and business products); and meta-webindexer (Meta AI search — Meta says allowing it helps its AI cite and link to you). A site-wide block of the training crawler:
User-agent: meta-externalagent Disallow: /
Keep meta-externalads allowed if you run Meta ads. Robots.txt changes can take up to 24 hours to apply, and Meta-ExternalFetcher and FacebookExternalHit may bypass robots.txt.
3. Check logs — and get advice if your content has licensing value. Look for unfamiliar bots in server or Cloudflare logs. If your content is licensed or syndicated, talk to a lawyer before October 30. Do not delete your Page in a panic: for most small businesses, Meta is a primary customer channel. Choose informed consent, not a hasty exit.
Bottom line: Meta removed the sentence tying business website content to AI training — a real win. But the crawling stays, the date stays, and terms can change again. Treat October 30 as a deadline to review.
