Meta has removed a specific clause from its commercial terms that previously stated the company could use content from designated domains to train its artificial intelligence products. The updated text no longer grants Meta the right to access data from web pages or URLs for the purpose of developing or maintaining its services. This change follows significant criticism from publishers who argued the original wording gave the tech giant a blanket licence to scrape news sites without explicit permission. Legal experts previously interpreted the language as allowing Meta to bypass standard scraping restrictions on news organisations.
The removal limits the scope of data Meta can legally harvest under these specific commercial agreements. It does not necessarily mean the company will stop training models on news content entirely, as other mechanisms might still apply. However, it removes the broad justification that publishers feared would lead to unauthorised access to their articles.
* The change applies specifically to the commercial terms page for business users.
* Legal analysts note the new wording is less ambiguous regarding scraping rights.
* Publishers hope this shift reduces the risk of their content being used without consent.
