Meta quietly deleted a line in its commercial terms that critics said gave it permission to scrape news sites for AI training data.
The removed language said Meta could use content from any domain, webpage, or URL to "provide, develop, improve, research, train, and maintain the integrity of" its products and services. Nieman Lab first flagged the wording, and Cornell information law professor James Grimmelmann said it read as a blanket license to hoover up published content, including journalism. After the backlash, Meta scrubbed the clause from its commercial terms page without issuing a public explanation for the change. This is not the first time the company has rewritten these terms under pressure.
Publishers have spent years fighting AI companies over whether scraping articles for training counts as fair use, and vague terms-of-service language only adds fuel to that fight. Quietly editing a legal document after getting caught, rather than clarifying intent up front, leaves open the question of what Meta's terms actually permit now.
Deleting a sentence is not the same as promising not to do the thing it described.