Meta has revised its commercial terms following wording on website crawling that raised concerns content from business websites could be used to train its AI models. The latest version drops the word “train” and limits the stated use of collected material to Meta’s advertising and business products.
Publishers have grown more careful about how AI companies handle their work. An Axios analysis of 102 major U.S. publishers found that 42% block AI training crawlers, while 25% block crawlers used for AI search.
Earlier Wording That Raised Alarms
Meta’s earlier terms allowed automated crawlers to collect content from websites linked to business accounts. The company stated that material could be used to “provide, develop, improve, research, train, and maintain the integrity of” Meta products.
Legal experts who spoke to Nieman Lab flagged how broadly that language could be read. They questioned whether it might give Meta permission to use content from publisher websites for AI training.
Meta pushed back on that reading. The company said the terms were meant for its advertising and business products, not to open a new path for scraping news sites to train AI.
What the Revised Terms Say
The updated version removes “train.” It now states that collected content can be used to “provide, develop, improve, research, and maintain advertising and business products.”
Crawling rights remain in the terms. Meta says site owners can still set scraping preferences through robots.txt. Legal experts note, however, that the revised terms do not explicitly require Meta to follow those instructions.
Broader Context for Publishers and AI Training
Meta has faced similar questions before about public content and AI training. The company uses public Facebook and Instagram posts for training in some regions. Users have reported problems when trying to object to their data being used.
Publishers are also seeking ways to get paid when AI companies use their work. Amazon has reportedly explored a marketplace where media companies could license content directly to AI firms. Similar licensing deals are becoming more common across the industry.