Human Approval
A mechanism that introduces human confirmation before an AI Agent performs sensitive operations, ensuring that operations are authorized and auditable.
Timeline (last 90 days)
Application-layer security controls should include three independent controls: input validation, human approval, and authorization, working together.
A reasonable approach is to require mandatory human confirmation at irreversible or high-impact operation checkpoints (such as merging to the main branch or deploying to production).
All Facts (2)
Application-layer security controls should include three independent controls: input validation, human approval, and authorization, working together.
50%UnverifiedA reasonable approach is to require mandatory human confirmation at irreversible or high-impact operation checkpoints (such as merging to the main branch or deploying to production).
50%