Amazon VP argues human-in-the-loop AI governance is not a quality standard

According to The Register, Eric Brandwine, VP at Amazon Security, stated that human-in-the-loop workflows fail as AI quality assurance metrics because humans are non-deterministic and inconsistent, no different than AI systems. Brandwine said humans and AI agents both make mistakes and produce variable outputs given identical inputs, but enterprises have millennia of experience managing human failure versus less than a decade with modern LLMs. Age of Product reports this position is now being debated at enterprise level as organizations rethink agentic governance models.