Meta has removed the word "train" from its streamlined commercial terms of service, set to take effect October 30. The change follows backlash after legal experts warned that the sweeping new terms could allow Meta to endlessly scrape websites linked on business pages — potentially sidestepping millions in licensing fees from news organizations.
The original language, flagged by Nieman Lab, appeared to require all business users to share a wealth of AI training data simply for operating a business page on the platform. Legal experts described the terms as "broad enough to drive a truck through," raising concerns that Meta could use the policy to build training datasets without compensating content creators.
In response, Meta provided a statement acknowledging the concerns: "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."
However, Meta will not add explicit language respecting robots.txt directives — a concession that experts say doesn't go far enough. The same legal scholars who raised the initial alarms note that the revised terms still leave significant ambiguity about what data Meta can collect from business pages and how it can be used for AI model development.
