Meta's terms reach your own domain, not just your Facebook page
Meta's updated terms reportedly let its crawlers take content from domains "associated with" an account and feed it into its AI products, with no opt-out. The fix publishers are landing on: price machine access separately from human access.
By PANONDA Newsroom

Foto: Bastian Riccardi · Pexels
What happened
Meta's updated terms reportedly authorize its web scrapers and bots to crawl domains "associated with" a user's account, and to feed what they find into Meta products — including its AI systems. The scope, as reported on November 2, 2026, covers journalism hosted on a publisher's own servers, not only posts uploaded to Facebook, Instagram or Threads. There is reportedly no opt-out in the terms as described.
That is the part worth sitting with. Most platform terms govern what you upload. This one, as reported, reaches what you publish somewhere else.
Why this is different from the last Meta terms story
We've covered the delay and the wording changes. This is the scope question, and it lands somewhere new.
The working assumption for most independent publishers has been that the platform relationship ends at the platform. Post to Facebook, accept Facebook's terms for that post, keep your site as the asset you control. If the reported scope holds, linking your domain to your Meta account is itself the grant — and leaving the platform later doesn't claw back what the crawler already took.
That reframes a decision a lot of newsletter and video operations made casually: which domain is listed on the business page, who set it up, and when.
What this story is not
It is not a confirmed enforcement record. The available reporting establishes the terms' reported scope. It does not confirm how Meta is applying the provision, which publishers have actually been crawled, or at what rate.
It is not confirmation that this is model training. Training, retrieval for AI answers, and generic "product improvement" are three different uses with three different legal temperatures. The reporting does not separate them.
It is not a payment story. No compensation mechanism, rate card or formula appears in the available material. Anyone quoting you a dollar figure for this is making it up.
And it is not a reason to blanket-block every bot tomorrow. Some of Meta's crawling is tied to how your links render and distribute on its apps. Cutting the crawler can cut the preview. Check before you flip a switch.
The shift underneath it
The durable response publishers are converging on isn't a lawsuit. It's a product decision: stop treating your site as one door.
Three tiers, priced separately. Human readers, free or paywalled as today. Licensed AI access, negotiated and paid. Unlicensed crawlers, metered or blocked. That is an operational change, not a legal one, and it doesn't require anyone's permission.
Separately, the National Newspaper Association joined a News/Media Alliance amicus brief in a Southern District of New York case where newspapers are trying to stop OpenAI from using local-news content without payment. The case is at summary judgment, where the court is asked to resolve copyright and fair-use questions. No ruling, no damages figure, no injunction yet. Litigation is slow. Your robots file is fast.
What to do this week
- Open your Meta Business account and list every domain associated with it. Include ones a contractor or former employee added. You cannot assess scope you haven't inventoried.
- Pull your server logs for the last 90 days and identify Meta's crawler user agents by volume. You need a baseline before you argue about anything.
- Decide which of those crawlers drive link previews and distribution, and which only ingest. Block the second group first, in a staged test, and watch referral traffic for two weeks.
- Write a one-page machine-access policy and publish it at a fixed URL. Human free, AI licensed, unlicensed metered. It costs nothing and it's the document any future negotiation starts from.
- Do not sign or renew a platform terms acceptance this month without reading the domain clause specifically. Have someone other than the person who clicks through read it.
- Keep a dated archive of the terms text as it exists today. Terms get edited quietly — we've already seen one word disappear.
- Skip the damages math. There is no confirmed price for this, and planning revenue against an unpublished formula is how budgets break.
Sources
Everything above was checked against these pages. Open them and see for yourself — that is why they are here.
Get the stories by email
What changes in AI for people who publish — sourced, and to the point.
No spam. Unsubscribe anytime.