Meta pulls crawling clause from commercial terms after backlash

Meta has removed a clause from its commercial terms of use that legal experts said could have authorized the company to scrape publishers' websites and use the content to train its AI models. The Verge reports that the provision is gone. The reversal came within days of the change drawing public criticism.

The clause, added in a recent update to the commercial terms, stated that Meta can "use any content from [a business account's] domains, web pages, or URLs to provide, develop, improve, research, train, and maintain the integrity of" Meta products and services. The language covered scraping and AI training in a single sentence. Legal experts interviewed by Nieman Lab called it a "remarkable provision."

Their concern was direct. Any publisher holding a Meta business account — a prerequisite for running pages, ads and messaging on Facebook and Instagram — could have been deemed to have consented to Meta crawling its entire website. The content could then feed the company's AI systems. Nieman Lab's experts read the terms as a potential blanket license.

What did Meta actually change?

Meta insists the answer is nothing. A company spokesperson said before the removal that the updated terms did not reflect a change in how Meta crawls the web. After the clause drew attention, the company issued a longer statement.

"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," Meta said in the statement. "It does not change how Meta uses web crawlers."

The company's account of the episode is that the language consolidated existing terms covering ads and business tools, and that readers mistook that consolidation for a new license. The removal of the clause, and the clarification embedded in the terms themselves, now stand as the record of Meta's position.

Why does one sentence in a terms-of-service document matter?

The stakes sit at the intersection of contract law and copyright law. Courts have so far given AI developers partial cover on the copyright front. Meta defeated a high-profile lawsuit over alleged copyright infringement last year. Other cases against AI companies, including Meta, continue to work through the courts.

A contractual grant changes that calculus. If a business account holder agrees in writing that Meta may "use any content" from its domains to train AI models, the company gains an argument that does not depend on fair use or on the outcome of copyright litigation. Publishers who accepted the terms to keep their Facebook pages would have handed over that grant without a separate negotiation.

That is why legal experts flagged the wording rather than treating it as boilerplate. One sentence in a commercial terms document could have functioned as a consent mechanism covering millions of business accounts.

What has changed for publishers now?

The clause is gone from the commercial terms. Meta's statement says the consolidated crawling terms applied only to ads and business products, and that the company has clarified that within the terms themselves. Meta maintains that its web crawling practices are unchanged from before the update.

For publishers, the immediate practical change is the absence of the language that triggered the alarm. The episode does not resolve the underlying questions about what content Meta's crawlers may collect from the open web, or how that content is used in AI training. Those questions remain the subject of active litigation.

What does the episode show about terms-of-service scrutiny?

The sequence unfolded quickly. Meta updated its commercial terms. Nieman Lab spoke with legal experts who identified the provision as remarkable and potentially sweeping. A Meta spokesperson said the terms did not change crawling behavior. The Verge then reported the provision's removal, accompanied by Meta's statement acknowledging the feedback.

The "We've heard you" framing signals that the company tracked the reaction and chose clarification over defense of the wording. For a company of Meta's scale, a terms-of-service edit is a low-cost concession compared with the alternative: a contractual clause that plaintiffs' lawyers could cite in the ongoing AI copyright cases.

Legal experts had connected the provision to exactly that litigation environment. Meta won one high-profile copyright suit last year, but the broader wave of cases against AI companies has not ended. A business-terms grant covering "any content" from account holders' websites would have given Meta a second, contractual line of defense — and removing it forecloses that argument before it is tested.

What happens next?

The removal closes this specific episode but not the larger conflict. Meta's crawlers continue to operate under the company's existing practices, which Meta says the update never altered. The remaining legal questions — what web content AI companies may collect, and on what terms — now return to the courtroom, where the pending cases against Meta and other AI developers will decide them. Publishers negotiating their relationship with Meta's platforms will be reading the commercial terms closely with each future update.