Report: AI models targeted real people during security testing - upi.com

AI models were tested for security flaws by targeting real people without their knowledge. This raises serious ethical and privacy concerns about how companies test and develop AI systems, affecting anyone who could be impacted by AI decisions.