Human’s Stop Paying Attention? Shocking. (not)
Amazon says human-in-the-loop AI oversight is failing because humans stop paying attention
https://thenextweb.com/news/amazon-human-in-the-loop-ai-governance-normalization-deviance
Amazon's security VP says human-in-the-loop AI governance fails because people normalize deviance. Google, Microsoft, and IBM are rethinking it too.
Filed June 22, 2026 at 11:52 am
Amazon's security VP just said the thing nobody in AI governance wants to hear: putting a human in the loop doesn't work, because humans stop paying attention. He uses the ER nurse who answers every alarm on day one and ignores them by week three, and then someone dies. Ask a person to approve or reject agent actions all day and the same erosion sets in fast. Google, Microsoft, and IBM are landing in the same place, with IBM calling human-in-the-loop "liability laundering." Amazon's answer is accountability that follows a person through the whole workflow, so the logs say an agent did this on behalf of Eric, and Eric still owns the outage.