---
title: "Autonomy level"
description: "How far an agent acts without a person between its steps, set by the deployer as a design decision rather than taken as a property of the model; one research scale names five levels by the user's…"
canonical: https://aigovernanceengineer.com/glossary/autonomy-level
author: "Jorge García Aibar"
license: "CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)"
doi: https://doi.org/10.5281/zenodo.22956197
version: "0.5.0"
updated: 2026-09-25
---

# Autonomy level

How far an agent acts without a person between its steps, set by the deployer as a design decision rather than taken as a property of the model; one research scale names five levels by the user's role, from operator to observer [1]. It is a registry field tied to a minimum control set, and raising it is a reviewed change.

- Developed in: [ch. 23, Autonomy is a design decision](https://aigovernanceengineer.com/bok/governing-agents#autonomy-is-a-design-decision)
- Chapters: [ch. 15, Deployment](https://aigovernanceengineer.com/bok/governing-deployment) · [ch. 17, Incidents](https://aigovernanceengineer.com/bok/incidents) · [ch. 23, AI Agents](https://aigovernanceengineer.com/bok/governing-agents)
- Contrast with: [Autonomy](https://aigovernanceengineer.com/glossary/autonomy)
- In the glossary chapter: https://aigovernanceengineer.com/bok/glossary#t-autonomy-level

## Sources

[1] "Levels of Autonomy for AI Agents" (K. J. Kevin Feng, David W. McDonald, Amy X. Zhang; arXiv 2506.12469; autonomy as a deliberate design decision separate from capability and operational environment; five levels by user role: operator, collaborator, consultant, approver, observer). arXiv. 2025-06-14 (v2 2025-07-28). https://arxiv.org/abs/2506.12469 (verified: primary)