Businesses cannot separately reject the updated terms while continuing to use the affected commercial services. Meta instead instructs organizations that disagree with the changes to stop using its products for business purposes.
The requirement could have significant implications for news publishers, many of which rely on Meta's platforms to distribute articles, reach audiences and promote their reporting.
Three legal experts who reviewed the agreement for Nieman Lab warned that its language could allow Meta to argue that publishers have already granted permission for their websites to be scraped and their content used for AI training.
Matt Topic, a media and intellectual property attorney at Loevy & Loevy, questioned the financial implications. Topic warned that the provision could give Meta access to material worth millions of dollars in potential licensing agreements. He also noted that it remains unclear whether courts would enforce such broad contractual permissions.
Meta disputes suggestions that the updated agreement changes how it collects information from websites.
The company announced the revisions earlier this month, describing them as an effort to consolidate existing policies covering commercial activities, advertising and AI-related services.
A Meta spokesperson told Nieman Lab that the company was combining provisions previously distributed across several agreements, including policies governing generative AI tools used in advertising.
In a separate statement to Ars Technica, Meta said “there is no change to how Meta uses web crawlers; any suggestion to the contrary is false.”
The company also maintains that website operators can continue restricting automated access through robots.txt files, which instruct web crawlers about which portions of a website they may access.
However, legal experts question whether Meta could eventually use the updated agreement to claim contractual permission to collect material even when website owners have attempted to block its crawlers.
The revised terms authorize Meta and companies acting on its behalf to use automated tools to retrieve, analyze and store information from websites associated with commercial accounts.
They also permit Meta to use collected material for developing, improving, researching and training its products. Although the provision does not explicitly mention artificial intelligence, Meta's broader product definitions include its AI offerings.
James Grimmelmann, a professor of digital and information law at Cornell Law School, told Nieman Lab that the agreement could grant Meta extensive permissions involving publisher content.
He warned that the company could potentially use material from news websites to train AI systems and distribute some of that information through its own products, creating competition with the original publishers.
Whether courts would uphold that interpretation remains uncertain. Legal experts emphasized that the agreement's enforceability has not been established, particularly where it intersects with copyright law.
The issue comes as publishers and AI developers continue fighting legal battles over whether copyrighted material can be used for model training without permission.
OpenAI faces lawsuits involving publisher content, while Meta has been sued by book publishers alleging that millions of copyrighted works were unlawfully obtained to train its Llama models.
Another provision attracting scrutiny allows Meta to create test credentials when accessing websites associated with business accounts.
Legal experts warn that these credentials could potentially allow automated systems to retrieve information from websites requiring registration or authentication.
Grimmelmann told Nieman Lab that the language could technically permit Meta to create accounts capable of accessing content unavailable to ordinary web crawlers.
Meta did not directly answer questions about whether the provision could be used to access subscription-only material or bypass paywalls.
An examination by Ars Technica found that similar permissions already existed in certain Meta agreements, although they were generally limited to specific commercial activities.
For example, Meta's seller agreement permits automated website access to support e-commerce listings, while its advertising policies allow website analysis for certain advertising services.
The new commercial agreement appears to extend those permissions across a broader range of products and business activities.
Benjamin Sobel, a law professor at the University of Wisconsin, compared the arrangement to the antitrust concept of tying, in which access to one service is conditioned on accepting another requirement.
Under Meta's revised terms, businesses seeking to continue using certain commercial features would also have to accept the company's website-access provisions.
Sobel compared the approach to Google's previous policies for AI-powered search, which required publishers seeking visibility in traditional search results to accept inclusion in AI-generated search features. Google introduced additional opt-out controls in August 2026.
Meta's changes also follow the introduction of Meta One, a paid subscription service that restricts how many external links certain Facebook and Instagram accounts can share without subscribing.
The combination has raised concerns that businesses could be required to pay for distributing links while simultaneously granting Meta permission to collect material from the websites those links reference.
However, Press Gazette reported that some news publishers are currently exempt from the subscription service's link restrictions.
For smaller publishers, the revised agreement presents a difficult choice because social media platforms remain important sources of readership.
Stephen Beale, founder and editor of The Steampunk Explorer, told Nieman Lab that Facebook plays a substantial role in connecting his publication with its audience.
Beale said he was considering closing his Meta accounts before October 30, despite the potential consequences for his publication's traffic.
The controversy leaves publishers weighing the benefits of maintaining their presence on Meta's platforms against the implications of accepting broader permissions over their website content.
Meta maintains that its existing crawling practices and robots.txt protections will remain unchanged. Legal experts caution, however, that the agreement's wording could become important in future disputes over AI training permissions and the use of copyrighted material.
This analysis is based on reporting from ars TECHNICA & Nieman Lab.
Image courtesy of The Verge.
This article was generated with AI assistance and reviewed for accuracy and quality.
