Meta Removes Terms Tied to Scraping News Sites for AI Training

Meta removes language on scraping news sites Meta has removed wording from its terms that experts said could allow the company to scrape news websites for material to train its artificial intelligence systems. The change follows attention to clauses in Meta’s terms of service that described how the company could use information available online. Experts interpreted that language as potentially giving Meta permission to collect and use content from news sites, even when the sites had not directly provided it for AI training. The wording is no longer in the terms. Its removal does not, by itself, establish whether Meta scraped news articles or used them to train its AI; the issue raised by experts was what the previous language appeared to permit. The change therefore addresses the terms’ description of Meta’s rights, rather than providing a statement about any specific scraping activity. Meta’s revised wording has drawn attention because it removes language that experts had viewed as relevant to using news content for AI training. What the removed terms concerned The language Meta removed related to the use of news websites to train artificial intelligence. It addressed the scraping of content from those sites, a practice that can involve collecting and processing material published online to help develop AI systems. The terms appeared in a section of Meta’s policies concerning how information could be used. Their wording specifically referred to scraping news sites for AI training, rather than describing a general change to Meta’s approach to news or its relationships with publishers. Removing the language means that this particular reference is no longer present in the terms. The deletion concerns the wording itself. It does not, on its own, explain why Meta made the change or establish what the company’s policy is now on collecting news content for training. Those questions require interpretation beyond the terms that were removed. The change nevertheless drew attention because the deleted passage had explicitly connected scraping news websites with the training of AI. Experts’ interpretation of the wording Experts said the wording of Meta’s terms allowed the scraping of news sites. Their interpretation concerned the language that had been included in the terms, rather than a new statement from Meta about whether scraping was permitted. The experts’ view was that the relevant wording allowed this activity. The terms had referred to scraping, and experts said that the language covered scraping news sites. This interpretation is part of the story of the wording Meta later removed. It describes how experts understood the terms before the change, without suggesting that they provided further details about their reasoning. In short, experts said the language allowed this scraping. The wording is important because Meta subsequently removed it from its terms. The change to the terms is covered separately. The change to Meta’s terms The wording at issue has been removed from Meta’s terms. The terms no longer include the language that referred to scraping news websites, following attention to the provision and its interpretation. The change is limited to the removal of that wording in the terms. The available information does not state why Meta removed the language. Nor does the removal, on its own, establish what Meta’s position is on scraping news sites or explain how its policies or practices may have changed. The change that can be confirmed is simply that the relevant language is no longer present in the terms. That distinction matters when describing the update: the text has changed, but the source does not offer a reason for the change or provide further detail about its significance. Any broader account of Meta’s intentions or the practical effect of the revised terms would go beyond what is stated here.