Meta has updated its commercial terms of service to allow automated scraping of business-linked website content for AI training, effective October 30, with no opt-out option. Legal experts warn this could enable Meta to train AI on news publishers' content without licensing fees, potentially undermining ongoing fair use debates with companies like OpenAI. The updated TOS permits Meta to use web crawlers to access, index, and analyze domains associated with business accounts, including news sites, to develop and improve its products.
While Meta claims no change in crawler usage and says organizations can block scraping via robots.txt, critics argue the terms could override such protections if courts side with Meta on AI training as fair use. James Grimmelmann of Cornell Law School warns Meta could deploy scraped content in competition with publishers. The move follows Meta's recent decision to charge users for posting links, suggesting a strategy to monetize link sharing then exploit the same links for AI training.
Experts call the provision remarkable for its potential to access content worth millions in licensing value.
Source: Ars Technica · Summarized by HeadlinesBriefing