Enum: RiskMeasurementChallengeEnum
Challenges that complicate AI risk measurement (Part 1 §1.2.1).
URI: nist_ai_rmf:RiskMeasurementChallengeEnum
Permissible Values
| Value | Meaning | Description |
|---|---|---|
| THIRD_PARTY_SOFTWARE_HARDWARE_AND_DATA | None | Risks related to third-party software, hardware, and data, |
| TRACKING_EMERGENT_RISKS | None | Identifying and tracking emergent risks and considering |
| AVAILABILITY_OF_RELIABLE_METRICS | None | Lack of consensus on robust and verifiable measurement methods |
| RISK_AT_DIFFERENT_LIFECYCLE_STAGES | None | Measuring risk at different stages may yield different results; |
| RISK_IN_REAL_WORLD_SETTINGS | None | Lab measurements may differ from risks that emerge in |
| INSCRUTABILITY | None | Opaque systems with limited explainability or interpretability, |
| HUMAN_BASELINE | None | AI systems intended to augment or replace human activity |
Slots
| Name | Description |
|---|---|
| challenge_kind | Which measurement challenge this represents |
In Subsets
Identifier and Mapping Information
Schema Source
- from schema: https://w3id.org/lmodel/nist-ai-rmf
LinkML Source
name: RiskMeasurementChallengeEnum
description: Challenges that complicate AI risk measurement (Part 1 §1.2.1).
in_subset:
- risk_and_harm
from_schema: https://w3id.org/lmodel/nist-ai-rmf
rank: 1000
permissible_values:
THIRD_PARTY_SOFTWARE_HARDWARE_AND_DATA:
text: THIRD_PARTY_SOFTWARE_HARDWARE_AND_DATA
description: 'Risks related to third-party software, hardware, and data,
including misalignment of risk metrics or methodologies between
developers, deployers, and operators.'
TRACKING_EMERGENT_RISKS:
text: TRACKING_EMERGENT_RISKS
description: 'Identifying and tracking emergent risks and considering
techniques for measuring them; impact assessment helps
understand context-specific harms.'
AVAILABILITY_OF_RELIABLE_METRICS:
text: AVAILABILITY_OF_RELIABLE_METRICS
description: 'Lack of consensus on robust and verifiable measurement methods
for risk and trustworthiness, and their applicability to
different use cases.'
RISK_AT_DIFFERENT_LIFECYCLE_STAGES:
text: RISK_AT_DIFFERENT_LIFECYCLE_STAGES
description: 'Measuring risk at different stages may yield different results;
some risks may be latent and increase as systems adapt and
evolve.'
RISK_IN_REAL_WORLD_SETTINGS:
text: RISK_IN_REAL_WORLD_SETTINGS
description: 'Lab measurements may differ from risks that emerge in
operational, real-world settings.'
INSCRUTABILITY:
text: INSCRUTABILITY
description: 'Opaque systems with limited explainability or interpretability,
poor documentation, or inherent uncertainty complicate risk
measurement.'
HUMAN_BASELINE:
text: HUMAN_BASELINE
description: 'AI systems intended to augment or replace human activity
require a human baseline for comparison, which is difficult to
systematize.'