Meta Platforms is set to implement new commercial terms on October 30, 2026, which will permit the company to crawl and collect data from websites linked to business accounts on its platforms, including Facebook, Instagram, and WhatsApp. The updated terms state that Meta may use this collected content to "develop, improve, research, train, and maintain the integrity of" its products, services, and features, explicitly including "AI at Meta."
This change has drawn scrutiny from legal experts and news publishers. Three legal professionals who reviewed the terms for Nieman Lab indicated that the language could allow Meta to scrape content from news sites for AI training without securing separate licensing agreements. This comes as many news organizations maintain business accounts with Meta to manage pages, run advertisements, and promote content.
Meta spokesperson Andy Stone stated that reports suggesting the company's updated commercial terms would allow it to scrape news sites are "false." Stone clarified that the update is intended to ensure Meta can retrieve accurate product information for on-platform ad and content recommendations, not to scrape news sites. He also noted that Meta's automated software will continue to respect robots.txt files, which website owners use to control web crawler access.
However, legal experts interviewed by Nieman Lab argue that the new commercial terms could provide Meta with legal grounds to bypass robots.txt instructions, as commercial contracts can supersede basic web blocking directives. Matt Topic, a media and intellectual property attorney at Loevy & Loevy, described the provision as "remarkable to bury in terms of service," suggesting it could give Meta access to material that would cost millions in a licensing market. He advised publishers to assume the language would be enforceable.
The updated terms also authorize Meta to "create test credentials" to access domains, web pages, or URLs. James Grimmelmann, a professor of digital and information law at Cornell Law School, suggested this could allow Meta to "sign up as a fake user" to circumvent registration walls on publisher sites. Meta declined to comment on whether test credentials would be used to bypass paywalls or other paid content.
Meta's current AI training practices involve a mix of publicly available data, licensed data, and information from its own products and services, including publicly shared Facebook and Instagram posts. The company has faced copyright litigation in the past regarding the alleged use of copyrighted books for training its Llama models.
Businesses using Meta's commercial products, such as those that manage Facebook pages, Instagram profiles, or WhatsApp channels, will be considered to have agreed to the new terms if they continue to use these services after October 30. There is no explicit opt-out option for the scraping clause; businesses that do not agree to the updated terms must stop using Meta's products for commercial purposes.