Meta has removed the word "train" from new commercial terms of service that take effect on 30 October, after criticism that they could let it use material from websites linked to business pages for AI training, Ars Technica reports. Legal experts had told Nieman Lab that the broad wording might let Meta avoid paying news organisations licensing fees by requiring businesses to hand over training data simply to run a page.
A Meta spokesperson confirmed to Ars that the company had not notified users of the change. In a statement, Meta said the consolidated crawling terms applied only to its advertising and business products, did not change how its web crawlers work, and did not override robots.txt preferences, adding that the update was not about scraping news sites. Experts quoted by Nieman Lab said the revisions did not go far enough and that publishers should remain concerned.
Source: Ars Technica