Meta updates terms that may widen its AI training access to publisher-linked content

The new rules take effect on October 30 and could extend Meta’s access to web content tied to businesses, beyond its existing licensing deals.

Meta has revised its business terms in a way that could let it crawl and use publisher-linked web content for AI training starting October 30, which matters because it may broaden the company’s access beyond its existing licensing deals. The policy says Meta can use automated tools on domains and pages tied to a business’s products or services, and that it may use material from those sources to develop, improve, research, train, and maintain its products and services. Meta says it is only consolidating its commercial and self-serve ad terms and that robots.txt still matters, but Nieman Lab’s reporting and legal experts suggest the wording could let Meta override blocking instructions and bypass licensing barriers. Cornell’s James Grimmelmann told Nieman Lab that the language looks like permission to train on content from news publishers and potentially use it in ways that compete with them. The company has already struck AI licensing agreements with CNN, Fox News, USA Today, and People Inc., and it recently introduced Meta One, a subscription that requires payment to include links in posts and reels.