Meta's new business terms may allow it to scrape news sites for free to train AI

Meta's new business terms, effective Oct 30, authorize Meta to use crawlers, bots, and spiders to access, index, cache, analyze, and scrape domains, webpages, or URLs tied to a business's products or services, and to pull updates periodically. Meta may use this content to develop, improve, research, and train its AI products. Meta spokesperson Daniel Roberts said this merely consolidates business and self-serve advertising terms, with no change in scraping methods, and that content owners can still express preferences via robots.txt. But legal experts told Nieman Lab the new terms could let Meta argue business terms override other blocking directives, letting it train models for free without licensing deals. For news publishers and online content operators, this means their site content could be used to train Meta's AI without a licensing agreement, and some content could be redeployed to compete with the publishers themselves. Cornell digital and information law professor James Grimmelmann said this effectively gives Meta a license to train AI on news publishers' site content. Businesses that disagree may lose access to all products or business uses. Nieman Lab also noted Meta last month launched the Meta One subscription, requiring businesses and creators to pay for links in posts and Reels—effectively charging users to post links while scraping the same content for free.