---
aliases:
  en:
  - trusted model monitoring
  - trusted monitoring protocol
concept_id: trusted_monitoring
preferred:
  en: trusted monitoring
relations:
  relations:
    broader:
    - control_protocol
    - oversight
    prerequisite:
    - control_protocol
    - oversight
    related:
    - untrusted_monitoring
status: active
---

# trusted monitoring

An AI-control protocol in which outputs or actions from an untrusted model are reviewed for suspicious or dangerous behavior by a monitor that is assumed not to intentionally subvert the protocol, often because the monitor is less capable, with suspicious cases handled by measures such as auditing, blocking, editing, or deferral.

## broader

- [control protocol](../terminology/control_protocol.txt)
- [oversight](../terminology/oversight.txt)

## prerequisite

- [control protocol](../terminology/control_protocol.txt)
- [oversight](../terminology/oversight.txt)

## related

- [untrusted monitoring](../terminology/untrusted_monitoring.txt)

[Corpus index](../index.txt)