PromptRiskDBThreat intelligence atlas

Autonomy/agency loss

AI Risk

"Autonomy/agency loss - Loss of an individual, group or organisation’s ability to make informed decisions or pursue goals."

Overview

A source-backed snapshot of this AI risk.

Techniques0Attack methods connected to this risk.
Mitigations0Defenses that may help with related attacks.
Records2Source records unified into this concept.

Risk profile

How the MIT AI Risk Repository categorizes this risk.

Domain5. Human-Computer Interaction
Subdomain5.2 > Loss of human agency and autonomy
Entity3 - Other
Intent3 - Other
Timing3 - Other
CategoryAutonomy
SubcategoryAutonomy/agency loss

Merged risk records

Source records unified into this canonical risk concept.

2 recordsView all →

MITRISK-Abercrombie2024-58.01.01 - Autonomy/agency loss

"Autonomy/agency loss - Loss of an individual, group or organisation’s ability to make informed decisions or pursue goals."

Domain5. Human-Computer InteractionSubdomain5.2 > Loss of human agency and autonomySourceA Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation HarmsYear2024

MITRISK-Li2025-66.01.03 - Autonomy / agency loss

"Loss of an individual, group or organisation’s ability to make informed decisions or pursue goals"

Domain5. Human-Computer InteractionSubdomain5.2 > Loss of human agency and autonomySourceA Closer Look at the Existing Risks of Generative AI: Mapping the Who, What, and How of Real-World IncidentsYear2025

Related techniques

Attack methods connected to this risk.

View all related techniques →
No linked attack methods. No AI attack method is connected to this risk in the current data.

Mitigations

Defenses that may help with related attacks.

View all mitigations →
No propagated mitigations. No defense is available through the connected attack methods.

Source evidence

Research source for this risk, when available.

View all sources →