Meta has updated its business terms to remove language that appeared to grant permission for training AI models on news sites, but legal experts argue the changes do not sufficiently restrict data collection. The new terms, which take effect on October 30, no longer explicitly state that Meta may use content from sites like The Verge for model training. Despite this adjustment, James Grimmelmann, a professor of digital and information law at Cornell Law School, told Nieman Lab that the revision fails to address core concerns. Grimmelmann noted that the terms still contain the phrase "associated with your products or services." He argued that this specific language allows Meta to scrape entire websites, meaning the initial permission to train AI on news content remains effectively intact through broader service definitions.