Skip to content
anythingai.space
AI Securityadvanced13 min · 10 steps

Human Approval for Actions

Make high-impact agent actions wait for the right person to approve the exact change — bound to its parameters, single-use and short-lived — so approval is a control, not theatre.

model's textapproveWHAT THE APPROVER SAW (MODEL TEXT)Routine password reset for E12345WHAT ACTUALLY RANadd_to_group(user=E12345, group=VPN-Contractors)Action agentread an injected ticketApproversponsoring managerDirectoryVPN-Contractorschanged
requesterrortool calldata
1 / 10
ProblemStep 1 of 10

The approver approved what the model told them

A contractor asks the IT assistant for VPN access. A malicious ticket the agent read earlier has injected an instruction, and the approval message the model writes says “routine password reset”.

The manager clicks Approve. What actually runs is add_to_group(E12345, VPN-Contractors) — remote network access. An approval that can be shown misleadingly, bypassed, replayed or self-approved is security theatre.

Watch outIf the model writes the approval screen, the attacker writes the approval screen.
What changed
  • +Approval message written by the model
  • ~Approver sees a different action than the one that runs

Lesson 16 of 22 in AI Security