Get the latest tech news
Meta ordered to stop training its AI on Brazilian personal data
This follows a similar pushback against AI training in the EU.
The decision follows a report published by Human Rights Watch last month which found that LAION-5B — one of the largest image-caption datasets used to train AI models — contains personal, identifiable photos of Brazilian children, placing them at risk of deepfakes and other exploitation. As reported by The Associated Press, ANPD told the country’s official gazette that the policy carries “imminent risk of serious and irreparable or difficult-to-repair damage to the fundamental rights” of Brazilian users. Meta received similar pushback from regulators in the EU causing the company to pause plans to train its AI models on European Facebook and Instagram posts.
Or read this on The Verge
