New · October 17, 2026
Meta backed down on scraping your site for AI training
Meta slipped a sweeping clause into its new commercial terms, legal experts called it a blank check to scrape publisher sites for AI training — and after a weekend of backlash, the clause quietly disappeared. Here's what actually changed, what didn't, and what to do about it.
What happened, in order
📝 The clause
Meta's updated commercial terms (effective Oct 30) said it could "use any content from [a business account's] domains, web pages, or URLs to provide, develop, improve, research, train, and maintain the integrity of" its products. Legal experts told Nieman Lab it looked like a "remarkable provision" — potentially covering any publisher with a Meta business account.
🔥 The backlash
Cornell digital-law professor James Grimmelmann and publisher groups (European Publishers Council, UK News Media Association, News/Media Alliance) went public. A Meta spokesperson first insisted the update "did not reflect a change" in how Meta crawls the web.
✂️ The retreat
On Oct 17, The Verge reported the provision had been removed from the terms page. Meta's statement: the crawling terms "only applied to our ads and business products… It does not change how Meta uses web crawlers." TechCrunch independently confirmed the removal the same day.
The honest read: what didn't change
Three caveats the headlines skipped:
- The word "AI" never appeared in the clause. "Train" in a terms-of-service sentence is not the same as an announced AI-training program — the interpretation came from legal experts, not from Meta's own framing.
- Meta says crawling practices are unchanged. Removing the clause takes away the written claim; it does not create a new barrier to scraping. The legal fight over scraping-for-training data is still being fought in courts and legislatures, not settled by one edit.
- robots.txt is still your real control. Meta reiterated that content owners can use robots.txt to set crawler preferences. If you haven't audited yours, that's the lever that actually works.
Does this change anything about Muse the agent?
Not this dispute. This story is about Meta's business/commercial terms for publisher websites — what it may do with content on sites tied to business accounts. It says nothing about what the Muse agent does with your messages, contacts, or connected accounts inside the app. Separate concerns live in separate reporting (e.g. the Oct 8 Malwarebytes Labs piece on Muse profiling), and this terms edit affects none of them.
3 prompts for site owners this week
Written for the practical work this story actually demands — no fear, just maintenance.
Audit your robots.txt
Show me my site's current robots.txt and user-agent rules. Explain in plain language which major crawlers (Google, Meta, OpenAI, Anthropic, Apple) I allow and which I block. Flag anything that accidentally lets training-oriented crawlers in, and propose a tighter version I can paste back — without breaking Google Search indexing.
Track the Oct 30 terms for me
Meta's commercial terms take effect October 30, 2026. Summarize the current version of the crawling/content-use section as of today, note anything that changed since the October 17 removal, and tell me in one paragraph what it means for a small publisher with a Meta business account. No speculation — only quote what the terms actually say.
Draft my crawler policy
I'm a small business with a website and a Meta business account. Draft a short, plain-English crawler policy for my site: which bots are welcome, which are restricted, and why. Include the robots.txt rules to match, and a one-line note I can put in my footer or privacy page so visitors know my stance on AI training data use.
Good to know
- The new commercial terms take effect October 30, 2026 for business users on Facebook, Instagram and WhatsApp — the Oct 17 edit happened before that date.
- This sits in the same week as Muse Spark Open's release (Oct 15): Meta is simultaneously giving away open weights and clarifying what web content it claims for training. Those are two different tracks — don't merge them.
- New here? Set up Muse in 15 minutes → · Get 1B free tokens →
← Muse Spark Open · Muse for Windows → · Where Muse is blocked · All news
Not affiliated with Meta. Figures: TechCrunch (Oct 17, 2026), The Verge, Nieman Lab via reporting, Meta spokesperson statement. Social claims that "Meta will no longer scrape creators' work" overstate the story — Meta says crawler practices are unchanged.