Meta removes AI training reference from website-crawling terms
The revised terms take effect on 30 October, but a US publishers’ group says they do not explicitly rule out training on website content.
Meta has revised upcoming terms governing access to business customers’ websites, deleting a reference to AI training following complaints from customers and media organizations, Gizmodo reported. The company says the provisions apply only to its advertising and business products.
The terms become effective on 30 October, with continued use of Meta’s commercial products counting as acceptance. According to PPC Land, the earlier version authorized automated access to websites connected to those products, including copying content, revisiting sites and using the material to train Meta’s products.
Meta told the Mercury News that the revision clarifies the intended scope rather than introducing a new restriction. The company also said its web crawlers would continue operating as before.
The News/Media Alliance, a group representing US publishers, questioned whether the changes provide sufficient limits. Spokesperson Sam Quigley told the Mercury News that removing training from the specified uses does not explicitly prohibit it. He also said the terms contain no commitment to honor robots.txt, a file website operators use to specify which pages crawlers can access.
