Meta announced a revision to its commercial Terms of Service that eliminates the word "train" after facing criticism over its use of scraped website content for artificial-intelligence purposes. The change follows a public outcry that the original language permitted broad data harvesting from business pages to improve Meta's products and services.
The earlier agreement gave Meta permission to use any material posted on business pages for a range of activities, including providing, developing, improving, researching, training, and maintaining product integrity. The updated version drops references to "any content" and the training function, instead allowing the company to periodically pull updates from business sites to support advertising and business tools, as well as to safeguard product security.
According to the new policy, the limited scraping is framed as a benefit to users, feeding Meta's research that helps advertisers create and deliver commercial content, inform users about products and services, and connect businesses with potential customers. The language emphasizes that the data collection is intended to improve ad delivery and user experience.
Legal experts note that the revised terms still lack any explicit commitment to respect robots.txt files, which website owners use to signal scraping preferences. University of Wisconsin law professor Benjamin Sobel told Nieman Lab that adding clear language about honoring robots.txt would address many concerns, but the current wording leaves the guidance voluntary and therefore ignorable.
Sobel explained that unless a preference is legally binding, it can be disregarded, allowing Meta to continue scraping sites that have attempted to block such activity. Other scholars suggest that Meta could rely on the enforceable Terms of Service to bypass robots.txt instructions, raising questions about the balance between contractual authority and web standards.
The policy shift has drawn criticism from news organizations worldwide, which are urging regulators to step in and prevent the contentious provisions from taking effect. Calls for oversight highlight worries that the revised terms could still enable large-scale data extraction without respecting established web-crawling norms.