The sentence in Meta's updated Commercial Terms that lets the company use content from business websites to "train" its products was still on Meta's published preview of those terms on October 18, 2026, a day after The Verge reported that Meta had removed it following criticism from publishers and legal experts. The terms take effect on October 30, 2026, and Meta has said it clarified in the text that the crawling provisions apply only to its ads and business products. No such limitation appears in the preview reviewed by PPC Land.

In Short

Meta wrote a line into the rules for every business on Facebook and Instagram saying it could use anything on those businesses' websites to train its products, and on October 17 The Verge reported that Meta had taken the line out. When PPC Land checked Meta's own preview of the new rules a day later, the line was still there, word for word, along with the permission to send crawlers to those websites. So businesses face two stories about the same contract, and the version that counts starts applying on October 30.

What The Verge reported

PPC Land reported on October 16 that Meta's new terms let it crawl business websites to train AI from Oct 30. The provision sat in Section 1 of the revised terms, renamed "Licences and rights you provide Meta," and had two sentences. The first granted the crawl. The second said what Meta could do with the results.

On October 17, The Verge reported that Meta was "updating its terms again" after the backlash, and that the second sentence "no longer appears on the commercial terms page." The publication quoted the passage it said had gone:

"… 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 provided under these Commercial Terms and any other applicable supplemental terms."

According to The Verge, Nieman Lab and James Grimmelmann, a professor of digital and information law at Cornell, had read that wording as an attempt to give Meta the right to use content scraped from sites to train its AI products for free and to supersede blocking instructions in a site's robots.txt file. The Verge did not say anything else in the paragraph had changed, and it did not quote a revised version of the clause.

What Meta's preview page shows

Meta publishes upcoming versions of the terms on a preview page, facebook.com/legal/commercial_terms_preview, under the notice "We're updating the Meta Commercial Terms. These updates will come into effect on 30 October 2026." PPC Land checked it on October 18. The UK English edition was captured at 20:44 UTC that day, nearly 30 hours after The Verge's report went up at 5:04 p.m. Central European Summer Time on October 17.

The sentence is there. So is the one before it:

"You authorise us (or any entity acting on our behalf) to use automated software (such as web scrapers, bots, robots, spiders, crawlers and user agents) and create test credentials to access, index, cache, analyse, collect, retrieve or crawl the domains, web pages or URLs (and any content available through them) that are associated with your products or services, or that are made available through your or users' use of Meta Company Products and services, and to periodically pull updates from such domains, web pages or URLs. 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 provided under these Commercial Terms and any other applicable supplemental terms."

PPC Land compared the October 18 capture with the copy of the updated terms used for its October 16 article. Apart from the site footer, the two are identical, word for word. Nothing in Section 1, Section 5 on AI features, Section 7 on updates or anywhere else in the document had been edited. The effective date at the foot still reads 30 October 2026.

How can both accounts be true? The supplied documents do not answer that. The Verge may have been reading a different language edition from the UK English preview, or a page Meta updated before the preview. Meta may be rolling the edit out in stages. Or the change described to The Verge may not yet have reached the published text. What can be said is narrower: on the evening of October 18, at least one official edition of the terms that takes effect on October 30 still contained the sentence The Verge reported as removed.

Meta's explanation, set against the text

Andy Stone, Meta's vice president of communications, and spokesperson Daniel Roberts had pushed back on the original reports, saying, according to The Verge: "There is no change to how Meta uses web crawlers; any suggestion to the contrary is false."

Roberts went further in a statement to The Verge on the morning of October 17:

"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. As always, content owners can use robots.txt to signal their preferences. This is not about scraping news sites, and these updates don't give us new rights that supersede those preferences."

Each of those claims can be checked against the preview. The crawl paragraph contains no reference to ads or business products. Its scope is any domain "associated with your products or services" or "made available through your or users' use of Meta Company Products and services," and its stated purposes include "train." Section 1 does describe "business, commercial and related services to help you create and deliver ads and other commercial content, provide users with information about your products and services, and connect you with users who may be interested in your products and services." That sentence was already in the October 16 version, though, and it governs the licence over content businesses upload to Meta, not the crawl. It is not a new clarification.

Nor does the preview mention robots.txt. Roberts said the updates "don't give us new rights that supersede those preferences." The contract text neither carves robots.txt out nor overrides it. It is silent.

Meta's own words to The Verge, then, describe a document that PPC Land could not find. That leaves the statement as a description of intent, which was also the status of Stone's denial on October 16.

Smaller discrepancies

The Verge attributes its "no change" line jointly to Stone and Roberts. Stone's October 16 post on X, covered in PPC Land's earlier article, was worded differently, opening with "This is false" and tying the update to product information for ad and content recommendations. The Verge does not say where each spokesperson's words were given.

The product framing has drifted too. On October 16 the explanation was products for sale on third-party websites. On October 17 it was "ads and business products." The second category is wider. A Facebook Page or a Meta Business Suite account is a business product with no product catalogue attached.

What the October 30 terms still do

If the preview is the text that applies, nothing in PPC Land's October 16 analysis has changed. Businesses authorise Meta, or any entity acting for it, to crawl, cache and collect content from covered sites and to create test credentials to reach them. Grimmelmann told Nieman Lab that test credentials could technically let Meta sign up as a fake user to get past registration walls; Meta declined to tell Nieman Lab whether it would use them that way.

Consent remains passive. Section 7 of the preview says that continuing commercial use after notice of an update means agreeing to it, and that "If you do not agree to the updated terms, you may stop all access or use of our Products for business or commercial purposes." There is no partial opt-out. The licence over uploaded content still requires businesses to warrant rights including "permission to display, distribute, create derivative works and deliver your Content," and Section 5 still makes them "solely responsible for and assume any and all risks associated with such AI features and any resulting Content."

Why it matters

For publishers, the stakes are the same as last week, plus a new uncertainty. Meta's crawler Meta-ExternalAgent uses one robots.txt token for both AI training and indexing, which makes it a mixed-use crawler: a site cannot allow one purpose and refuse the other. A clause in which the site owner expressly licenses training would weaken any later argument that a robots.txt block reflected the owner's wishes. A public report that the clause is gone, sitting beside a preview that still contains it, makes it harder for a business to know what it is agreeing to by posting on its Page after October 30.

For advertisers, Meta's stated reason, accurate product data for ads, lines up with product browsing switching on in existing ads on October 28. Advertisers relying on catalogue feeds were never dependent on the training sentence, and nothing in the preview narrows the crawl to product pages either.

Three questions remain open. Which version of the text did The Verge see, and is it live in every language and region? Where is the ads and business products limitation Roberts says Meta added? And will the preview match it before October 30? Twelve days remain before continued use of a Page or an ad account counts as acceptance.

Timeline

  • March 6, 2024 - Effective date of the previous Meta Commercial Terms, which contain no crawling clause.
  • Early October 2026 - Meta notifies business users of updated Commercial Terms taking effect on October 30.
  • October 15, 2026 - Nieman Lab publishes legal experts' readings that the new terms could allow scraping of news sites.
  • October 16, 2026 - PPC Land reports the new terms let Meta crawl business websites and train AI from October 30; Andy Stone calls the reports false.
  • October 17, 2026 - The Verge reports Meta has removed the "train" sentence; Daniel Roberts says Meta clarified in the terms that crawling applies only to ads and business products.
  • October 18, 2026 - Meta's preview of the terms, UK English edition captured at 20:44 UTC, still contains the sentence and no ads-only limitation.
  • October 28, 2026 - Product browsing switches on for existing Meta ads.
  • October 30, 2026 - Updated Meta Commercial Terms take effect.

Summary

Who: Meta and spokespeople Andy Stone and Daniel Roberts; The Verge, which reported the removal; and every business using Meta's products for commercial purposes, including news publishers and advertisers.

What: The Verge reported on October 17 that Meta removed the sentence in its updated Commercial Terms allowing it to use content from crawled business websites to "train" its products, and Roberts said Meta had clarified in the terms that crawling applies only to ads and business products. Meta's preview of the terms, checked by PPC Land on October 18, still contains the sentence unchanged, the crawl authorisation and test credentials language, and no ads-only limitation.

When: The Verge reported on October 17, 2026. The preview was captured on October 18, 2026. The terms take effect on October 30, 2026.

Where: The Meta Commercial Terms apply to commercial use of Meta Company Products worldwide. The preview reviewed is the UK English edition at facebook.com/legal/commercial_terms_preview.

Why: Legal experts read the sentence as a free licence to train AI on news sites that could be argued to outrank robots.txt. Meta's public account of the terms and the text it has published now diverge, twelve days before continued use counts as acceptance.