Meta removes clause from commercial terms allowing AI training on websites

The inclusion of external domains and links tied to corporate profiles within material usable by Meta platforms was rendered void on October 17, 2026. The change directly affected press outlets, news sites, and publishing companies with a presence on the social network.

The multinational's rectification followed legal analyses indicating that the wording opened the door to massive digital data extraction without financial compensation for content owners. The company removed the text after complaints from the technology and media sectors.

The contractual clause that sparked the debate

The previous version of the commercial terms stated that Meta could use any content from domains, websites, or links associated with professional accounts to provide, develop, improve, research, train, and maintain the integrity of its products and services. The specific inclusion of research and training within this contractual document raised concerns among companies with a presence on its networks.

The use of automated tools for digital data extraction is the standard procedure for building learning bases for modern language models. Including these permissions in mandatory professional account clauses represented a direct path to accessing third-party content libraries.

Warnings from the legal and publishing sectors

Digital law specialists consulted by Nieman Lab described the previous wording as an extraordinary provision. James Grimmelmann, a professor of digital and information law at Cornell University, warned that the contractual language could be interpreted as express authorization to collect material from news websites for free.

Grimmelmann also warned that consent to these commercial terms could contractually override robots.txt exclusion directives. This circumstance would have rendered the technical instructions used by publishers to block access to automated crawlers practically ineffective.

Meta's response and rectification

Meta spokespeople initially rejected the idea that the update to the rules represented a change in the behavior of their crawling systems. Andy Stone, vice president of communications at the multinational, and Daniel Roberts, a company spokesperson, called claims suggesting a modification in the technical operation of their robots false.

In a statement sent to The Verge, Daniel Roberts confirmed the definitive removal of the questioned fragment. The spokesperson clarified that the unified terms applied only to the operation of advertising and commercial products. Roberts insisted that website owners can continue to use the robots.txt file to set their preferences and remarked that Meta's rules do not override those technical decisions.

This contractual controversy coincides with other ongoing litigation for the company. Although Meta overcame an intellectual property infringement lawsuit last year, the company remains involved in other legal proceedings regarding the use of protected content for training algorithms.