trust
The trust layer of agentic systems — what makes an agent's output verifiable, what makes its actions reversible, and the gap between what vendors claim and what users can check.
Recommended starting points
All pieces filed in trust
Essay
Black Box AI: When You Can't Audit the Stack, How Do You Trust It?
The agentic era has made the trust problem worse. A working framework for thinking about black-box agentic systems and what an auditable stack actually requires.
Black Box AIauditability
Authors active in this beat
Related topics