Meta modified its business terms to delete language indicating the company could freely train artificial intelligence models on news websites. The company removed these specific provisions after facing significant pushback from the publishing industry.
Original terms stated that Meta might use any content from websites linked to a business account to research, develop, and train its products. New business account terms are scheduled to take effect on 30 October 2026.
Critics argue the updated version still permits Meta to use scraped data to develop any business product. James Grimmelmann said, "This does nothing to restrict the ‘associated with your products or services’ [language] that would let Meta scrape entire sites."
Matt Topic suggested that the word 'develop' is broad enough to encompass 'train'. Legal analysts contend the changes are insufficient to address ongoing concerns.
The terms might enable Meta to claim its commercial agreements override other blocking instructions such as robots.txt. Meta maintains that the new terms are not intended to scrape news content for free.
Daniel Roberts, a Meta spokesperson, stated that content owners can use robots.txt to signal their preferences. Roberts also stated the update is not about scraping news sites and does not give Meta new rights that supersede robots.txt preferences.
