Why Amazon hates 'human-in-the-loop' AI governance
We like to think we are all very good at our jobs, and we have high opinions of ourselves, he explained during a phone interview with The Register. “But when you actually get down to it, humans are not terribly consistent,” Brandwine said. Humans, like AI agents and systems, are non-deterministic.
- ▪We like to think we are all very good at our jobs, and we have high opinions of ourselves, he explained during a phone interview with The Register.
- ▪“But when you actually get down to it, humans are not terribly consistent,” Brandwine said.
- ▪Humans, like AI agents and systems, are non-deterministic.
Opening excerpt (first ~120 words) tap to expand
(function() { let windowUrl = window.location.href; windowUrl = windowUrl.substring(windowUrl.indexOf('?') + 1); let messageElement = document.querySelector('.shareableMessage'); if (windowUrl && windowUrl.includes('code') && windowUrl.includes('expires')) { messageElement.style.display = 'block'; } })(); Security Why Amazon hates 'human-in-the-loop' AI governance VP Eric Brandwine explains people aren't all that great, actually Jessica Lyons Jessica Lyons Published sat 20 Jun 2026 // 15:25 UTC Humans tend to be “a little bit precious about humans,” according to Eric Brandwine, distinguished engineer and VP at Amazon Security. We like to think we are all very good at our jobs, and we have high opinions of ourselves, he explained during a phone interview with The Register.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at theregister.
Discussion
0 commentsMore from theregister
