Skip to content

Enum: TrustworthinessPropertyType

Properties contributing to AI trustworthiness, enumerated in Clause 5.15 (robustness, reliability, resilience, controllability, explainability, predictability, transparency, fairness and bias-related properties).

URI: iso22989:TrustworthinessPropertyType

Permissible Values

Value Meaning Description
robustness None Ability to maintain performance under varied or adversarial conditions
reliability None Consistent intended behaviour over time under stated conditions
resilience None Ability to recover acceptable behaviour after disruption or failure
controllability None Property of allowing authorised humans to intervene in system behaviour
explainability None Property of producing explanations of system behaviour that are intelligible ...
predictability None Property of behaviour being anticipatable given known inputs and state
transparency None Property of disclosing meaningful information about the system to interested ...
fairness None Property of avoiding inappropriate or harmful discrimination across groups
bias_mitigation None Property of identifying and reducing unwanted bias in data, models or outcome...
accountability None Property of having identifiable parties answerable for system behaviour and o...
privacy nist_ai_rmf_common:TrustworthinessCharacteristicEnum#PRIVACY_ENHANCED Property of respecting personal data and individual privacy expectations
safety nist_ai_rmf_common:TrustworthinessCharacteristicEnum#SAFE Property of not causing unacceptable risk of harm to people, property or envi...
availability None Property of being accessible and usable on demand by authorised entities (Cla...
integrity None Property of safeguarding accuracy and completeness of data, models and output...
authenticity None Property of being able to verify the origin and identity of inputs, models an...
security None Property of preserving confidentiality, integrity and availability of the AI ...
usability None Property of being effectively, efficiently and satisfactorily usable by the i...
quality None Aggregate property reflecting how well the AI system meets stated and implied...
fault_tolerance None Property of continuing to operate correctly in the presence of component faul...

Slots

Name Description
trustworthiness_property_type Which trustworthiness property is being characterised

In Subsets

Identifier and Mapping Information

Schema Source

LinkML Source

name: TrustworthinessPropertyType
description: Properties contributing to AI trustworthiness, enumerated in Clause 5.15
  (robustness, reliability, resilience, controllability, explainability, predictability,
  transparency, fairness and bias-related properties).
in_subset:
- trustworthiness
from_schema: https://w3id.org/lmodel/iso22989
close_mappings:
- nist_ai_rmf_common:TrustworthinessCharacteristicEnum
rank: 1000
permissible_values:
  robustness:
    text: robustness
    description: Ability to maintain performance under varied or adversarial conditions.
  reliability:
    text: reliability
    description: Consistent intended behaviour over time under stated conditions.
    narrow_mappings:
    - nist_ai_rmf_common:TrustworthinessCharacteristicEnum#VALID_AND_RELIABLE
  resilience:
    text: resilience
    description: Ability to recover acceptable behaviour after disruption or failure.
    narrow_mappings:
    - nist_ai_rmf_common:TrustworthinessCharacteristicEnum#SECURE_AND_RESILIENT
  controllability:
    text: controllability
    description: Property of allowing authorised humans to intervene in system behaviour.
  explainability:
    text: explainability
    description: Property of producing explanations of system behaviour that are intelligible
      to relevant audiences.
    narrow_mappings:
    - nist_ai_rmf_common:TrustworthinessCharacteristicEnum#EXPLAINABLE_AND_INTERPRETABLE
  predictability:
    text: predictability
    description: Property of behaviour being anticipatable given known inputs and
      state.
  transparency:
    text: transparency
    description: Property of disclosing meaningful information about the system to
      interested parties.
    narrow_mappings:
    - nist_ai_rmf_common:TrustworthinessCharacteristicEnum#ACCOUNTABLE_AND_TRANSPARENT
  fairness:
    text: fairness
    description: Property of avoiding inappropriate or harmful discrimination across
      groups.
    narrow_mappings:
    - nist_ai_rmf_common:TrustworthinessCharacteristicEnum#FAIR_WITH_HARMFUL_BIAS_MANAGED
  bias_mitigation:
    text: bias_mitigation
    description: Property of identifying and reducing unwanted bias in data, models
      or outcomes.
    narrow_mappings:
    - nist_ai_rmf_common:TrustworthinessCharacteristicEnum#FAIR_WITH_HARMFUL_BIAS_MANAGED
  accountability:
    text: accountability
    description: Property of having identifiable parties answerable for system behaviour
      and outcomes.
    narrow_mappings:
    - nist_ai_rmf_common:TrustworthinessCharacteristicEnum#ACCOUNTABLE_AND_TRANSPARENT
  privacy:
    text: privacy
    description: Property of respecting personal data and individual privacy expectations.
    meaning: nist_ai_rmf_common:TrustworthinessCharacteristicEnum#PRIVACY_ENHANCED
    close_mappings:
    - iso29100:PrivacyPrinciple
    - nist_ai_600_1:GaiRiskCategoryEnum#DATA_PRIVACY
    related_mappings:
    - iso27001:SecurityControl
  safety:
    text: safety
    description: Property of not causing unacceptable risk of harm to people, property
      or environment.
    meaning: nist_ai_rmf_common:TrustworthinessCharacteristicEnum#SAFE
  availability:
    text: availability
    description: Property of being accessible and usable on demand by authorised entities
      (Clause 5.15.3 context).
  integrity:
    text: integrity
    description: Property of safeguarding accuracy and completeness of data, models
      and outputs (Clause 5.15 Note).
    close_mappings:
    - nist_ai_600_1:GaiRiskCategoryEnum#INFORMATION_INTEGRITY
  authenticity:
    text: authenticity
    description: Property of being able to verify the origin and identity of inputs,
      models and outputs (Clause 5.15 Note).
  security:
    text: security
    description: Property of preserving confidentiality, integrity and availability
      of the AI system and its data (Clause 5.15 context).
    close_mappings:
    - nist_ai_600_1:GaiRiskCategoryEnum#INFORMATION_SECURITY
    related_mappings:
    - iso27001:InformationSecurityManagementSystem
    narrow_mappings:
    - nist_ai_rmf_common:TrustworthinessCharacteristicEnum#SECURE_AND_RESILIENT
  usability:
    text: usability
    description: Property of being effectively, efficiently and satisfactorily usable
      by the intended users (Clause 5.15 Note).
  quality:
    text: quality
    description: Aggregate property reflecting how well the AI system meets stated
      and implied needs (Clause 5.15 Note).
  fault_tolerance:
    text: fault_tolerance
    description: Property of continuing to operate correctly in the presence of component
      faults (Clause 5.15.4 context).