Meta has removed language from its commercial terms of use that appeared to allow the company to use content from business customers' websites to train AI models. The clause covered "any content" from domains, web pages or URLs linked to a business account. For publishers and companies that maintain a Meta business account, the reversal removes an immediate legal risk, but leaves open how platform crawling rules will be written next.
Why the updated wording drew objections
The disputed wording said Meta could use such content to provide, develop, improve, research, train and maintain the integrity of its products and services. Legal experts interviewed by Nieman Lab described it as a remarkable provision with unusually broad scope. Their reading was that any publisher holding a Meta business account could have its site treated as available for scraping and subsequent model training under that language.
The concern centered on mechanics rather than a confirmed scraping program. A business account typically connects a company domain, pages and URLs to Meta systems for ads and business products. The consolidated crawling terms placed that connected web presence inside the same document that governs product development and training. Meta did not present evidence of expanded crawling, and a company spokesperson said the update did not reflect a change in how Meta crawls the web.
The episode arrived against continued copyright litigation around AI training data. Meta defeated a high-profile lawsuit over alleged copyright infringement last year, while other cases against AI companies, including Meta, continue. Publishers therefore read broad contractual phrases as potential rights to collect text, images and pages at scale. The Verge reported that the provision has now been taken out of the terms after the criticism.
What the reversal means for business websites
For companies that advertise on Meta or use its business products, the practical effect is a return to the prior ambiguity rather than a new guarantee. Meta stated that the consolidated crawling terms applied only to ads and business products and said it had clarified that point in the terms themselves. It added that the change does not alter how Meta uses web crawlers. Marketing and media teams keep the same operational setup, without an added contractual basis for site reuse.
The limits of the statement deserve attention during vendor and legal review. The company framed the issue as clarification of existing ads and business-product terms, not as a policy on AI training data, crawler behavior or opt-out controls. The source material provides no crawler names, no exclusion mechanism, no effective dates and no list of products covered by training. Contracts should therefore be checked for what crawling is permitted, for which product, and what happens to collected content afterward.
The marker to watch is the next published version of the commercial terms and any accompanying crawler documentation. If the removed sentence stays out and Meta specifies which crawlers serve ads measurement and which serve model development, confidence for publishers will increase. If broad development and training language returns in another section, businesses with valuable archives will need to reassess account settings, access controls and licensing positions.
