Meta drops AI training from the website-crawling clause in its new business terms

The permission to crawl any website tied to a business account stayed in, and it takes effect on October 30.

By Think Facility · October 18, 2026 · Meta AI policy

Meta has taken the word "train" out of the website-crawling clause in its new business terms. Two Internet Archive copies of its terms preview page, one saved on the morning of October 16, 2026 and one on the afternoon of October 17, show the edit.

The permission to crawl stayed in. The terms take effect on October 30, 2026, and they cover any business that runs ads, lists products or manages a Page on Meta's apps.

Takes effect: October 30, 2026
Covers: using ads, selling or listing products, managing a Page
Word removed: train
Still allowed: crawl the domains, web pages, or URLs associated with your products or services
Current terms: effective March 6, 2024

What changed in the clause

The first version was still on Meta's page early on October 16. It let Meta use content from those sites for pretty much anything it builds.

A large white sign with a blue and white thumbs-up icon above a black base reading facebook and 1 Hacker Way, in front of dark green trees — The sign at 1 Hacker Way, Meta's campus in Menlo Park, California. Photo: Minette Lontsie, CC BY-SA 4.0, via Wikimedia Commons

Meta may use any content from such domains, web pages, or URLs to provide, develop, improve, research, train

From Facebook

By the afternoon of October 17 that sentence was gone. Now the purpose sits inside the crawling sentence: Meta can pull updates from those sites "in order to provide, develop, improve, research, and maintain advertising and business products" and keep its products secure. Train isn't on the list anymore.

Most of the paragraph didn't move, though. Meta (or anyone acting for it) may still send scrapers, bots and crawlers, and even create test credentials, to reach any domain "associated with your products or services". It can still index and cache what it finds. And "research" and "develop" both survived, while the new text never says how far "business products" stretches.

Who objected, and what Meta said

Publishers moved first. The News/Media Alliance put out a statement from its president and CEO, Danielle Coffey, after Nieman Lab reported the clause. She called the terms "outrageous and probably illegal" and asked whether a website run by Taylor Swift or Shutterstock would be open to Meta's crawlers just because the owner held a Meta business account.

every publicly-hosted word, song, and image is up for grabs if Meta gets away with this

News/Media Alliance Statement on Changes to Meta Commercial Terms of Use | News/Media Alliance

Andy Stone answered for Meta on X at 10:28 am Eastern on October 16. The archived page still carried the old wording that morning.

Andy Stone@andymstoneOct 16, 10:28 am ET

This is false. There's no change to how Meta uses web crawlers. As always, content owners can use robots.txt flags to indicate their preferences. This update is designed to make sure we’re able to pull accurate information about products for sale on third-party websites when making on-platform ad and content recommendations, not about scraping news sites.

View the post on X

His post and the terms don't quite line up. Stone talks about product listings. The terms cover any domain tied to a business account, and neither version of the clause mentions robots.txt.

What it means if you run a website or a business Page

This reaches well past newsrooms. A corner shop with a Facebook Page counts, by Meta's own definition, because managing a Page sits on its list of business purposes right next to running ads. There's no opt-out box either. Keep using Meta's products after notice and you've agreed; if you don't agree, the updates section says, you may stop using them.

So what can a site owner actually do? Meta points to robots.txt, and its crawler page lists which bot does what. Meta-ExternalAgent is the one Meta says crawls for "training foundation AI models". Meta-ExternalAds handles advertising. You can block each by name, and changes can take up to 24 hours to stick.

Two crawlers may skip the file anyway. One fetches links a user asked for. The other, FacebookExternalHit, can ignore it during security checks.

What the terms leave open

The terms don't say they override a robots.txt block, and they don't say they respect one. I'd like to see that written into the contract before October 30, because an X post isn't one.

Until October 30, the current commercial terms apply. They date from March 6, 2024 and have no crawling paragraph at all.