Meta has revised its business terms to eliminate explicit language that suggested it could freely train its AI models on news websites. However, legal experts argue that the changes do not sufficiently address ongoing concerns regarding content scraping. James Grimmelmann, a professor at Cornell Law School, noted that the revised terms still contain vague language that could permit Meta to scrape entire sites.
Initially, the updated terms, effective October 30, indicated that Meta could utilize any content from websites linked to business accounts for “developing” and “training” its products. This raised alarms among legal experts, who warned that such language could allow Meta to override existing restrictions, such as those specified in the robots.txt file, which traditionally guides web crawlers on what content they can access.
In response to significant backlash from the publishing industry, Meta removed the specific mention of using scraped content for training its AI. Nevertheless, experts remain skeptical. Matt Topic, a media and intellectual property attorney, pointed out that the revised terms still permit Meta to use scraped content for developing any “business product,” a term that lacks a clear definition and could encompass AI-related products.
Despite these concerns, Meta maintains that its updated terms are not intended to facilitate the scraping of news content without permission. A spokesperson for the company emphasized that content owners can still use robots.txt to communicate their preferences regarding content access, asserting that the changes do not grant Meta new rights that would override these preferences.
As the debate continues, the implications of Meta’s terms on the relationship between tech companies and content creators remain a critical issue in the evolving landscape of digital rights and AI development. For further insights, see The Verge.
