Meta has adjusted its business account terms to remove wording that suggested it could freely train its AI models on news sites, including The Verge. However, legal experts told Nieman Lab that the revisions do not go far enough. James Grimmelmann, a professor of digital and information law at Cornell Law School, said the changes do nothing to restrict language permitting Meta to scrape entire sites based on content “associated with your products or services.”
The original terms, set to take effect October 30th, initially appeared to grant Meta permission to train its AI on news sites. The company stated it “may use” such material, according to the report.
Why it matters
The dispute concerns whether publishers’ content can be used to train AI models without clear restrictions. Legal experts suggest the amended terms still leave room for broad data collection, raising ongoing questions about publisher rights.
Who should care
News publishers, media organizations, and legal observers tracking how platform terms govern AI training on web content.