Meta is preparing to implement an update to its commercial terms on October 30, stating that the company or entities acting on its behalf may use automated software—including crawlers, bots, and test credentials—to access domains, pages, and links associated with business accounts’ products or services. The new wording allows the content to be indexed, cached, analyzed, collected, and retrieved, in addition to being used to provide, develop, improve, research, and train Meta products.
According to the published text, the terms do not give business accounts a clear option to reject these permissions while continuing to use Meta’s services. Instead, coverage indicates that companies that do not agree may have to stop using the platform’s services.
Why does this matter?
The significance of the update lies in the fact that it could shift the permission from narrow areas related to advertising or e-commerce to general wording covering content available through websites linked to business accounts. The potential impact is particularly significant for news organizations, which are already engaged in disputes over whether using their articles to train artificial intelligence models without a license constitutes fair use.
Three legal experts who spoke to Nieman Lab warned that the enforceability of these terms in court is unclear, but said the wording could be used to argue that Meta obtained permission to crawl publishers’ websites and use the content to train artificial intelligence products. James Grimmelmann, a professor of information and digital law at Cornell Law School, said the text could allow AI training on any content from their websites, and possibly the use of part of that content in products that compete with publishers on Meta’s platforms.
What does Meta say has changed?
Meta describes the amendment as a process of “simplifying” and consolidating provisions that had been distributed across different agreements, including terms governing the use of generative artificial intelligence in advertising. A company spokesperson said that the way crawlers are used has not changed and that any suggestion to the contrary is incorrect. The company also confirmed that content owners can still manage crawler access using a robots.txt file.
However, an Ars Technica comparison of the old and new wording identified new or broader language, including a reference to “test credentials.” Experts believe this phrase could permit automated access to content protected behind paywalls, although Meta pointed to the previous Vendor Agreement, which contained similar wording, but did not mention it in its post announcing the terms update.
Open questions for publishers
The central dispute remains related to the breadth of the permission and how it conflicts with robots.txt instructions, particularly if courts conclude in the future that training artificial intelligence on certain types of content constitutes fair use. Meta currently says it does not plan to use the terms as a basis for ignoring instructions to block crawling, but publishers are questioning whether its position will remain the same as the legal environment changes.
Experts also pointed to the possibility that the terms could represent an example of contractual “tie-in,” in which companies are required to accept broad permissions to access content as a condition of using another service. This could lead news organizations and companies to reassess their reliance on Meta as a source of traffic, although the article does not establish that widespread departures from the platform have occurred so far.