The story changed, but the published preview didn’t. The Verge reported on October 17 that Meta had pulled the sentence from its updated Commercial Terms allowing it to use content crawled from business websites to train its products. When PPC Land checked Meta’s official preview a day later, the sentence was still there, unchanged, with the crawl authorisation and the word “train” intact.
What Meta said vs what the text says
Meta communications VP Andy Stone and spokesperson Daniel Roberts pushed back on the first wave of coverage. Roberts told The Verge that the consolidated crawling terms “only applied to our ads and business products.”
But the preview PPC Land reviewed has no ads-only limitation in the crawl paragraph. The scope extends to any domain connected to your products or services, or made available through use of Meta’s commercial and consumer tools. The stated purposes still include training. And robots.txt, which Roberts said content owners can still use, is not mentioned in the contract text at all.
That leaves a simple problem: Meta’s public statement describes a document the published preview does not show.
Why this matters for advertisers and publishers
For publishers, the legal exposure is not just theoretical. Cornell professor James Grimmelmann read the clause as a licence to use scraped content for AI training and to sit above robots.txt blocking. Meta’s crawler, Meta-ExternalAgent, uses one robots.txt token for both AI training and indexing, so a site cannot allow search visibility while refusing training.
For advertisers, there is an extra wrinkle. Meta has said the crawl supports accurate product data for ads, and product browsing switches on for existing ads on October 28. But the preview does not narrow the crawl to product pages, and catalogue advertisers were never dependent on this sentence.
Consent is passive. Section 7 says continued commercial use after notice means agreement; the only opt-out is to stop all commercial use. There is no partial button.
What to do before October 30
- Check the live text. Open Meta’s commercial terms preview in your region and language, not just the media coverage.
- Audit robots.txt. Decide whether you can block Meta’s mixed-use crawler without hurting indexing.
- Get legal sign-off. If you run a Page, catalogue or ad account, continued use after October 30 counts as acceptance.
The practical rule here is trust but verify: when a platform’s PR and its published contract diverge, the contract is the version that governs.
Source: PPC Land
