Meta Removes Controversial AI Training Language Over Website Scraping Concerns

Meta revised its commercial terms to remove explicit language permitting the use of external content for AI training, responding to publisher and legal concerns over broad data scraping rights. While the original wording threatened to allow the collection of copyrighted material from business-affiliated websites for model development, the updated terms narrow the scope to activities strictly related to advertising and business products. This change aims to clarify that Meta’s intent is to support its commercial ecosystem rather than grant unrestricted access for artificial intelligence research. The removal of specific "training" references does not alter Meta’s actual web crawling practices or its ability to access content for improving business services. Publishers remain concerned that broad clauses allowing content collection for "developing and improving" products could still implicitly cover AI training, even without explicit mention. The reliance on standard technical controls like robots.txt highlights the ongoing tension between automated data collection norms and site owners’ efforts to restrict access, leaving legal boundaries ambiguous. This development is crucial for the open_data community as it illustrates the fragility of data access permissions within private platform contracts. It underscores how corporate terms can significantly impact the availability and licensing of public web data, even when technical scraping remains possible. The episode serves as a warning that legal clarity in commercial agreements is essential for predictable data usage, particularly as the line between service improvement and AI model training continues to blur in regulatory discussions.

Source: itechpost.com
Published on 2026-10-20

Meta Removes Controversial AI Training Language Over Website Scraping Concerns

Related news