I-HITL: Inalienable Human-in-the-Loop
arXiv cs.AI · 27 Oct 2026, 03:52 AEST
The concept of inalienable human-in-the-loop (I-HITL) is introduced to enhance AI safety by ensuring explicit human approval for actions taken by AI agents. Experiments reveal vulnerabilities in existing HITL systems, prompting the need for robust protocols like I-HITL.
Key facts
- Human oversight is a pillar of AI safety.
- An I-HITL protocol ensures that an agent can initiate an action only with explicit human approval, even when the human interacts with a system compromised by the agent.
- We report on experiments with a popular agentic wallet platform in which raising an agent's spending cap requires HITL approval via emailed one-time passcodes (OTPs).
- We show that an I-HITL protocol requires a trusted entity - a trusted computing device (or equivalent) - and characterize which such devices can and cannot realize an I-HITL protocol.
Read aloud
The concept of inalienable human-in-the-loop I-HITL is introduced to enhance A I safety by ensuring explicit human approval for actions taken by A I agents. Experiments reveal vulnerabilities in existing HITL systems, prompting the need for robust protocols like I-HITL. Human oversight is a pillar of A I safety. An I-HITL protocol ensures that an agent can initiate an action only with explicit human approval, even when the human interacts with a system compromised by the agent. We report on experiments with a popular agentic wallet platform in which raising an agent's spending cap requires HITL approval via emailed one-time passcodes OTPs. We show that an I-HITL protocol requires a trusted entity - a trusted computing device or equivalent - and characterize which such devices can and cannot realize an I-HITL protocol.
Sources in this cluster
- I-HITL: Inalienable Human-in-the-Loop
arXiv cs.AI · 27 Oct 2026, 03:52 AEST