Meta AI Scraping: Updated Terms Still Leave Door Open for News Site Scraping

Meta announced a revision to its business account terms on October 22, 2026, that removed a clause explicitly permitting the company to train AI models on content scraped from partner sites. The change was a response to backlash from publishers who argued that the original wording allowed Meta to override standard web-scraping controls such as robots.txt. However, legal scholars point out that the new language still grants Meta a broad right to "use any content" from sites associated with a business account to "develop, research, and train" its products, leaving the possibility of training models on news content open.

The Original Clause and Publisher Concerns

When Meta first rolled out its business terms in early 2026, the document included a sentence that read: "Meta may use any content from the websites associated with a business account to develop, research, and train Meta’s products." Publishers interpreted the word "train" as a direct permission to harvest articles for AI training. The clause was seen as a legal lever that could supersede the technical restrictions set by a site’s robots.txt file, which traditionally tells crawlers what they may or may not index.

The Revision and Its Ambiguities

The updated terms, effective October 30, 2026, excise the word "train" from the clause. The revised wording now states that Meta may use content to "develop, research, and train" its products, but the removal of the explicit training permission was framed as a clarification rather than a restriction. Legal experts argue that the term "develop" is sufficiently broad to encompass training activities, especially when Meta’s own products include AI services. Matt Topic, a media and intellectual property attorney, notes that "develop" is not a defined term in the contract, and that Meta is likely developing AI business products.

Meta’s Stance on Robots.txt

In a statement to The Verge, Meta spokesperson Daniel Roberts reiterated that content owners can use robots.txt to signal their preferences. Roberts emphasized that the update "is not about scraping news sites, and these updates don’t give us new rights that supersede those preferences." The company maintains that the revised terms do not grant it a blanket right to bypass standard web-scraping controls.

Legal Expert Analysis

James Grimmelmann, a professor of digital and information law at Cornell Law School, cautions that the revised terms still allow Meta to scrape entire sites under the umbrella of "associated with your products or services." He points out that the clause does not restrict the scope of content that can be used for development. The legal community remains divided on whether the removal of the word "train" materially changes Meta’s ability to harvest news content.

Potential Impact on Publishers

Publishers that rely on robots.txt to block crawlers may find that Meta’s broad development clause could still be invoked to justify data collection. If Meta interprets "develop" to include training, publishers could face continued exposure of their content to AI model training pipelines. The lack of a clear definition for "business product" adds uncertainty to how the clause will be applied in practice.

What to Watch Next

The next steps for publishers and regulators will likely involve clarifying the legal scope of Meta’s development clause. If Meta continues to argue that "develop" covers training, publishers may need to seek stronger contractual protections or pursue legal action to enforce robots.txt preferences. Regulatory bodies may also examine whether Meta’s terms violate antitrust or data-protection laws by granting an unfair advantage in AI training.

Bottom Line

Meta’s updated business terms remove an explicit training permission but leave a broad development clause that could still enable AI training on scraped news content. Publishers must monitor how Meta applies the clause and consider legal avenues to enforce their robots.txt signals. The situation remains fluid, and further clarification from Meta or regulators could reshape the landscape for news site scraping in the near future.

Related coverage

Explore more on this topic