Meta removes terms that raised concerns over website scraping and AI training

Meta has removed language from its commercial terms that appeared to allow it to scrape websites linked to business accounts and use their content to train its artificial intelligence (AI) models. The change follows criticism from legal experts and news publishers, who warned that the terms could give the company broad access to website content without separate licensing agreements.

The provision was introduced as part of an update to Meta’s commercial terms, which stated:

Meta may use any content from such domains, web pages, or URLs to provide, develop, improve, research, train, and maintain the integrity of, the Meta Company Products, services, and features provided under these Commercial Terms and any other applicable supplemental terms.

Meta removed the word “train” from the revised terms and limited the permitted uses to advertising and business products. However, legal experts told Nieman Lab that the revised wording still left room for broad website scraping. Meta subsequently removed the disputed provision, according to The Verge.

What raised concerns?

Nieman Lab reported on October 15 that the proposed terms could allow Meta to scrape entire websites associated with business accounts, including news publishers’ websites. These accounts allow organisations to manage their Facebook pages, Instagram profiles and other Meta services. The terms were scheduled to take effect on October 30, with no option for business accounts to opt out. Businesses that disagreed with the changes were told to stop using Meta’s products for business or commercial purposes.

The proposed terms also authorised Meta to create “test credentials” to access websites. James Grimmelmann, a professor of digital and information law at Cornell Law School, said this could allow the company to register as a fake user to get past registration barriers. Meta declined to answer Nieman Lab’s questions about whether these credentials could bypass paywalls or access paid content.

The revised wording did not explicitly address whether Meta would respect website instructions in robots.txt files, which tell automated crawlers which parts of a website they should not access. Meta said website owners could continue using these instructions to manage scraping.

“We’ve heard you and want to clarify: the crawling terms we consolidated only applied to our ads and business products, and we’ve clarified that in the terms themselves. It does not change how Meta uses web crawlers. As always, content owners can use robots.txt to signal their preferences. This is not about scraping news sites, and these updates don’t give us new rights that supersede those preferences,” a Meta spokesperson said.

Legal experts question the revised wording

Grimmelmann told Nieman Lab that the revised terms did not resolve his concerns. He argued that the phrase “develop…advertising and business products” remained broad enough to cover AI training.

Benjamin Sobel, a law professor at the University of Wisconsin, also cautioned that robots.txt expresses a website owner’s preferences but does not, by itself, create an enforceable legal right to prevent scraping.

Previously, its terms for generative AI advertising allowed it to crawl websites for the specific purpose of optimising advertisements. Extending similar permissions to broader advertising and business products raised concerns that the company could gain access to content beyond that limited purpose.

Publishers call for regulatory scrutiny

The proposed changes drew opposition from publishers’ groups. The European Publishers Council said it would seek scrutiny from European regulators over the terms’ implications for copyright, contractual fairness and market power. The UK News Media Association called for government intervention, while the US-based News/Media Alliance described the changes as “outrageous and probably illegal.”

The dispute comes amid wider concerns about technology companies using publicly available, copyrighted material to train AI models without licensing agreements. Meta has previously faced copyright litigation over its AI training practices, although it won a high-profile case involving copyrighted books in 2025. Other legal disputes involving AI companies remain ongoing.

Read more:

• Meta deploys AI tools to detect child abuse ads, says it actioned 5.3 million content items in India
• Meta continued to run AI-generated CSAM ads weeks after India investigation