Amazon VP Eric Brandwine argues human-in-the-loop AI governance is not the optimal oversight model

According to The Register, Eric Brandwine, distinguished engineer and VP at Amazon Security, stated that humans are non-deterministic and inconsistent like AI systems, and that human-in-the-loop governance is not necessarily the gold standard for managing automated and agentic systems. Brandwine noted that organizations have millennia of experience with human failure but less than a decade with modern AI systems.