What changed in the terms
Meta amended its business terms after harsh criticism from the publishing industry. The initial version, effective from October 30, stated that the company could use any content from sites linked to a business account to develop its products, conduct research, and train its models. According to The Verge, the wording suggested that Meta could train AI on that content.
After publishers' protests, Meta removed the reference to training on collected content. The revision is the central fact of the matter: the offending word was deleted from the text, but the rest of the contractual framework remained in place.
The legal experts' objections
For the experts cited by Nieman Lab, the change does not solve the problem. James Grimmelmann, a professor of information law at Cornell Law School, argues that the clause on content associated with products or services is not limited—the one that in his view would allow Meta to collect entire sites.
Matt Topic, a lawyer specializing in media and intellectual property, notes that the revised version still allows the use of collected content to develop a commercial product. Because the term does not appear to be defined, and Meta develops commercial AI-based products, according to the legal expert the word “develop” would be broad enough to include training. The criticisms are expert statements reported by the source, not judicial findings.
Meta's position and the open issues
Meta denies that the new terms are intended to collect journalistic content for free. Spokesperson Daniel Roberts reiterated to The Verge that content owners can use the robots.txt file to signal their preferences and that the update does not grant the company new rights that override those preferences.
One point raised earlier by the experts remains to be clarified: whether the business terms can be invoked to override sites' blocking instructions. The matter also involves Google only because the source cites robots.txt as a tool historically used to indicate to crawlers, including search engines, what they may visit. Meta, for its part, has not provided a definition of the term “develop” nor clarified how it will interpret the remaining clause. Publication of the new terms is set for October 30 according to The Verge, but the concrete effect on the handling of journalistic content cannot be verified with the available material.
