Amazon Challenges the Effectiveness of Human-in-the-Loop AI Oversight

Amazon’s VP of Security, Eric Brandwine, has expressed skepticism about the reliability of human-in-the-loop oversight in AI governance. Speaking to The Register, Brandwine pointed out that humans tend to lose focus and are inconsistent, which undermines the belief that human oversight is the best standard for AI security. This perspective challenges a widely held assumption among companies about the effectiveness of human-in-the-loop systems.