Meta Retracts Controversial AI Scraping Provision After Expert Backlash
Meta removed broad data scraping terms following legal scrutiny over AI training permissions.
1Meta Reverses Policy Update
Meta officially removed a controversial provision from its commercial terms of use on October 17. The update had previously suggested that the company could scrape content from external domains, including news sites, to train its AI models. This swift retraction followed significant public and expert backlash regarding data ownership rights.
2Legal Experts Raise Alarms
Legal scholars, including Cornell professor James Grimmelmann, identified that the deleted provision would have allowed Meta to bypass robots.txt instructions. Experts argued this would grant the company free access to copyrighted material, effectively superseding standard site-owner blocking protocols. This interpretation triggered immediate scrutiny regarding the legality of Meta's data collection practices.
3Scope of the Provision
The now-deleted text stated that Meta could use content from URLs to provide, develop, and maintain its products. Critics highlighted that this language was broad enough to encompass news sites and proprietary web content. The clause aimed to consolidate terms for ads and business products but faced intense criticism for overreach.
4Meta's Official Stance
Company spokesperson Andy Stone clarified that the update did not represent a change in web crawling behavior. Meta maintains that the consolidated terms were intended only for ads and business products. The company insists that content owners retain the ability to use robots.txt to signal their preferences against automated data scraping.
5Broader Industry Data Practices
The incident highlights growing tensions over data usage in the AI race. Companies like xAI currently use public posts from platforms like X to train chatbots without explicit user consent. Meta similarly scrapes publicly shared content from users over 18, fueling ongoing debates regarding privacy, copyright, and the ethics of AI.
6Precedent in Legal Battles
Meta previously defeated a high-profile copyright infringement lawsuit last year, yet legal challenges persist. Multiple proceedings against AI companies remain active in courts, focusing on how data is harvested for training. These cases underscore the volatile legal environment surrounding the use of internet-wide data for large-scale machine learning model development.
7Comparison With WeTransfer
The Meta situation mirrors recent industry trends, such as the case of WeTransfer in July 2025. WeTransfer faced public backlash after updating terms to include machine learning improvements, eventually removing the language entirely. These instances demonstrate a recurring pattern where tech firms attempt to quietly expand data usage rights before facing pressure.
8Future of Data Governance
This development signals that tech giants face heightened accountability regarding their terms of service. As AI training demands more data, the friction between platform policies and content creators will likely intensify. Stakeholders must now monitor how companies balance their aggressive AI development goals with increasing demands for transparency and user control.