What we know

Meta has updated its business terms to remove language that implied it could freely use content from news sites for training its AI models. Despite this change, legal experts say Meta may still be able to scrape news sites for AI training purposes because copyright law and the enforceability of website terms of service remain unclear. The company’s revised terms no longer explicitly permit AI training on news content, but no legal barriers have been definitively established to prevent such scraping.

Why it matters

Meta’s revision responds to concerns from publishers and journalists about unauthorized use of their content for AI training. Removing explicit permission from its terms signals an attempt to address these issues. However, legal experts highlight that ongoing ambiguities in copyright law and the limits of terms of service enforcement mean Meta could continue to use news content without clear legal consequences.

What is still unknown

It remains unclear how effective Meta’s updated terms will be in practice at preventing unauthorized scraping of news sites. No legal rulings or enforcement actions have clarified the boundaries for AI training data use in this context. The scope of Meta’s AI training data sources, the accuracy of its models, and any plans for further policy changes have not been disclosed. No independent assessments of Meta’s compliance or the impact of these changes are available.