Meta has updated its commercial terms again after strong pushback from publishers and digital law experts. The company removed a passage that, according to critics, could authorize the use of content from sites linked to business accounts — including news outlets — to train its AI models.
The removed language stated that Meta could use any content from associated domains, pages or URLs to "develop, improve, research, train and maintain the integrity" of its products. The new commercial terms take effect on October 30, 2026.
What the company says
In a statement to The Verge, spokesperson Daniel Roberts said the company heard the criticism and clarified that the consolidated crawling rules apply only to advertising and business products. According to Meta, there is no change in how it uses web crawlers, and content owners can continue using robots.txt to signal preferences. Andy Stone, vice president of communications, reiterated that the update is not aimed at scraping news sites.
Experts such as James Grimmelmann, professor of digital law at Cornell, and Nieman Lab had warned that the original wording could be interpreted as broad permission to train AI on publishers' material without licensing, potentially overriding robots.txt instructions.
Context
The controversy began with the consolidation of commercial terms announced by Meta. News publishers are already involved in similar disputes with other AI companies over the use of their content for training. The removal of the word "train" and Meta's clarification respond directly to this pressure.
Legal questions remain about whether contractual terms can prevail over technical controls and about the exact scope of the remaining permissions.
Sources
Transparency: This content was created, edited or reviewed with the assistance of artificial intelligence. The information was cross-checked with public posts on X and sources available on the internet. Consult the original sources to verify the full context.
By GeekikiBot