Meta Pulls Contested AI Scraping Clause From Commercial Terms After Experts Sound Alarm

Meta has removed language from its commercial terms that critics said could have opened the door to free AI training on scraped news content. The revision followed sharp criticism of the Meta scraping terms, which experts and Neiman Lab interpreted as granting the company broad permission to use material from domains, pages and URLs for developing, improving, researching and training artificial intelligence systems.

What the deleted clause said

The disputed passage was part of a consolidation of Meta’s crawling terms. It referred to content from domains, pages and URLs, and it listed AI-related activities including development, improvement, research and training. The removed text also referenced Meta Company Products, services and features, as well as supplemental terms.

Critics argued that the wording could let Meta train AI models on scraped content without payment or permission. They further warned that the clause might be read as overriding robots.txt, the longstanding convention that lets site owners signal which parts of their sites should not be crawled.

Why experts objected

James Grimmelmann, a Cornell University professor of digital and information law, interpreted the clause as creating a free right to train AI on the material. Neiman Lab reached a similar conclusion, describing the language as permitting scraping. The objections centered on the difference between a routine terms update and what critics saw as a substantive expansion of Meta’s rights over third-party content.

Meta’s initial response

Meta initially pushed back. Andy Stone, the company’s vice president of communications, and spokesperson Daniel Roberts denied that any change had been made to how Meta’s web crawlers operate. Their position was that the consolidation did not alter crawler behavior.

That denial did not quiet the criticism. The company then removed the disputed language entirely and issued a follow-up statement alongside the terms update.

The follow-up clarification

In a statement to The Verge, Roberts said Meta had heard the feedback and clarified the scope of the terms. He said the consolidated crawling terms applied only to ads and business products, not to the broader use of content that critics had feared.

Roberts also said robots.txt remains available for content owners to express their preferences, and that the update grants no new rights that supersede those preferences. According to Meta, the matter is not about scraping news sites.

What remains unchanged

Meta says its crawler behavior has not changed. The company maintains that robots.txt preferences still apply, and that the terms update was a scope limitation rather than an expansion of rights. The exact date of the terms change was not given, and the full text of the revised terms was not provided.

Whether any sites were actually scraped, or whether legal challenges followed, is not established. What is clear is that the episode put a spotlight on how commercial terms, crawler preferences and AI training data intersect — and on how quickly a quiet terms update can become a public dispute.