Meta has removed commercial-terms language that legal experts said could authorize the company to use content from business-account websites for AI training. The deletion addresses a specific concern for publishers: whether operating a Meta business account could give the company rights over their websites beyond the content they deliberately provide to Meta.
The Verge reported the removal on October 17, 2026, following backlash over the wording. Meta’s explanation is narrower than the interpretation that prompted the controversy. The company says the clause concerned ads and business products, did not change its crawler behavior, and did not grant new rights overriding website owners’ robots.txt preferences.
That distinction matters. Meta has reversed the disputed wording, but the available reporting does not establish that it reversed an actual scraping practice or stopped a previously authorized AI-training operation.
Why the Clause Alarmed Publishers
The removed passage said Meta could use content from specified domains, web pages, or URLs to provide, develop, improve, research, train, and maintain the integrity of its products, services, and features.
The concern was the combination of that broad description of website content with permission to “train” Meta products. Legal experts interpreted it as potentially reaching beyond material a business intentionally uploaded to Facebook or supplied for an advertising campaign.
According to TechCrunch’s account of the dispute, legal experts interviewed by Nieman Lab thought the provision could permit Meta to scrape the website of a publisher with a Meta business account and use the content to train AI models. The Verge identified Cornell digital and information law professor James Grimmelmann among those reading the language broadly.
Those were interpretations of the clause’s potential legal effect, not findings that Meta had exercised those rights. The concern also extended to whether accepting the terms could supersede instructions blocking crawlers through a website’s robots.txt file.
For publishers, the operational question was straightforward: could maintaining a business relationship with Meta undermine restrictions they had placed on access to their own sites?
That question was relevant beyond organizations buying advertisements. Meta’s commercial terms cover business or commercial uses including advertising, selling products, developing apps, managing a Page, managing a Group for business purposes, and using measurement services. A publisher reviewing its exposure therefore had reason to examine the terms even if its immediate concern was Page management rather than an ad campaign.
The disputed interpretation would have connected those platform activities to permissions over content hosted elsewhere. Removing the wording matters because businesses no longer have to assess that particular passage as part of the terms.
Meta Says It Never Claimed the Broader Rights
Meta rejects the premise that the clause expanded its web-crawling rights.
Spokesperson Daniel Roberts told The Verge that “the crawling terms we consolidated only applied to our ads and business products,” and said Meta had clarified that scope in the terms themselves. He also said the update did not change how Meta uses web crawlers.
On the robots.txt question, Roberts said content owners could continue using the file to signal their preferences. According to Meta, the updates did not give the company new rights that superseded those preferences, and the provision was not about scraping news sites.
This is more specific than a general denial that Meta trains AI on web content. The company’s response concerns what this particular contractual update meant and whether it altered existing crawler practices. It should not be read as a blanket statement about every source Meta uses for AI development.
The disagreement therefore has two parts. Critics saw wording broad enough to authorize website-content use for training, potentially despite crawler-blocking instructions. Meta says the intended scope was ads and business products, with no corresponding change to crawling or robots.txt treatment.
The deletion resolves the immediate wording dispute without establishing which interpretation would have prevailed legally. Meta’s clarification is its stated position, not an independent determination of the removed clause’s effect.
What the Removal Does Not Establish
Three separate questions sit behind the controversy: what a contract permits, what a crawler accesses, and what content enters a training process. Evidence about one does not automatically answer the others.
The reporting establishes that disputed language was removed and that Meta supplied a narrower explanation. It does not establish whether Meta used any publisher content under that wording before changing it.
It would therefore be inaccurate to describe the development as Meta admitting that it scraped publishers’ sites against their robots.txt instructions. It would also go beyond the evidence to say the removal proves no such collection occurred. Neither conclusion follows from a terms edit.
The timing needs similar care. October 17 is the date of The Verge’s report, followed by TechCrunch’s coverage. The available evidence does not identify the exact time Meta edited the terms page.
TechCrunch’s article also points to The Verge for the removal and carries Meta’s explanation. It provides additional coverage of the controversy, but it should not be treated as an independent technical investigation into crawler behavior.
These limits do not make the change insignificant. Businesses make decisions based on contractual language as well as technical behavior. Removing a passage that experts considered expansive can materially change the question their legal and publishing teams need to review, even when the company insists its underlying practices never changed.
What Businesses Should Check Before Changing Accounts
For a publisher deciding whether a Meta business account now grants blanket permission to train on its website, the most defensible answer is that the disputed wording has been removed, and Meta explicitly says this update did not create new rights overriding robots.txt preferences.
That is narrower than saying a business account has no implications for content permissions. Meta’s commercial terms still describe licenses for content shared, posted, uploaded, or made available on or in connection with its products. Separate product terms may also apply. The removal should not be treated as cancelling those existing arrangements.
Sources
- The Verge reported the removaltheverge.com
- TechCrunch’s account of the disputetechcrunch.com
- commercial termsfacebook.com
