Meta's New Terms Permit AI Training on Publisher Content
Meta's new terms allow the company to scrape publisher sites for AI training, sparking concerns over unpaid use of news content.
Tech
Meta announced that its updated platform terms, effective Oct. 30, will let the company crawl and scrape websites linked to businesses that use its services, including online news publishers. The language in the terms grants Meta permission to use this external content for purposes such as development, improvement, research, and training of artificial‑intelligence systems.
The change was highlighted in a report by Nieman Lab, which noted that the wording could allow Meta to harvest news articles without paying the publishers. The policy applies to sites that Meta deems "associated" with advertisers or other commercial partners on its platforms.
Industry observers say the move could intensify debates over data ownership and compensation for news outlets whose material may be used to fuel large‑scale AI models.
Why It Matters
U.S. and European news organizations could see their articles used to train AI without compensation, raising legal and ethical questions about data rights.
