Meta’s new terms likely allow AI to scrape website content

Meta, the parent company of Facebook, Instagram and WhatsApp, is introducing major changes to its commercial terms. Effective October 30, 2026, the revised terms authorise Meta to use automated software to access websites associated with businesses, collect their content and use it to develop, research, train and improve its products and services.

The inclusion of the word “training” has led experts to believe that the new terms could also cover the use of website content to train Meta’s AI models. The terms apply to organisations that manage Facebook pages, advertise on Meta’s platforms or use Meta Business Suite for commercial purposes. News organisations are not exempt.

US legal experts, commenting in the media, have warned that the revised terms could give Meta broad access to online news content, allowing it to copy substantial amounts of material without payment or specific conditions.

What the new terms say

The first section of the revised terms states: “You authorise Meta (or any entity acting on Meta’s behalf) to use automated software, such as web scrapers, bots, robots, spiders, crawlers and user agents, to create test credentials, access, index, analyse and collect information from websites, including regularly updated information.”

The authorisation applies to two types of websites: domains associated with a user’s products or services, and domains encountered through the use of Meta’s platforms.

The same section says Meta may use any content from those domains, webpages or URLs to provide, develop, research, train and secure its products, services and features. The reference to training raises concerns that content from these websites could be used to train Meta AI models.

The terms define “business or commercial purposes” broadly. These include sending advertisements or marketing messages, selling products, developing apps, managing pages or business groups, and using products such as Ads Manager or Meta Business Suite. The terms apply regardless of the type of organisation, provided it uses Meta’s platforms for such purposes.

Continued use means acceptance

Under Section 7 of the terms, users who continue using Meta’s products for business purposes after receiving notice of the changes will be deemed to have accepted the revised terms. No separate mechanism for rejecting the changes has been provided. Those unwilling to accept the terms must stop using Meta’s products for commercial purposes altogether.

Meta notified users of the changes through Business Suite in early October. The revised terms have also been published on Facebook’s website, where the page states that they will take effect on October 30, 2026.

What Meta says

Meta communications executive Andy Stone wrote in a post on X that the purpose of the update was to ensure the accurate collection of information about products listed for sale, not to scrape news websites. He also said the update would not change how Meta uses web crawlers.

In statements to several US media outlets, another Meta spokesperson said content owners could use robots.txt to control access to their websites.

The robots.txt file is a standard text file used by websites to specify which automated crawlers are allowed to access particular parts of a site and which are not.

Legal experts raise concerns

US legal experts, however, argue that the language of the revised terms is broad enough to cover news organisations’ websites. Almost every major news outlet has at least one Meta business account, they note.

They fear Meta could argue that acceptance of the new terms overrides restrictions imposed through robots.txt. Such an interpretation could potentially allow the company to avoid the need for content-licensing agreements with news publishers.

One expert has suggested that the content covered by such licensing arrangements could be worth many millions of dollars.

Meta has rejected these concerns, maintaining that there will be no change to its use of web crawlers.

Implications for Bangladesh’s news media

Facebook is a major source of traffic for online news outlets in Bangladesh. Almost all news organisations operate Facebook pages to distribute their stories, while many also advertise on the platform. As a result, the revised terms could affect Bangladeshi publishers as well.

The terms also specify how commercial disputes will be resolved. Disputes involving businesses outside the United States are to be handled by courts in California or Ireland. This means that a Bangladeshi publisher seeking to challenge Meta over the terms could have to pursue legal action outside the country.

Publishers can block Meta’s AI crawler, meta-externalagent, by configuring their websites’ robots.txt files. However, blocking facebookexternalhit, the crawler responsible for generating previews of links shared on Facebook, could significantly undermine their ability to distribute news through the platform.

Disabling that crawler could prevent Facebook from displaying link previews, including headlines and images, when news stories are shared. This would make links less useful and could reduce the visibility of publishers’ content on the platform.

Bangladeshi news organisations therefore face a difficult choice: restrict automated access to their content or risk losing some of the features that make Facebook an important channel for reaching readers.