Technical
Data risk
Risks from data quality, provenance, representativeness, and privacy.
Definition
A principal risk category covering issues related to data quality, provenance, freshness, representativeness and labelling that can produce biased, inaccurate or noncompliant AI outputs; it requires measures such as data provenance checks, representativeness assessments, privacy‑preserving techniques and cross‑border data safeguards.
Related Terms
Dataset Provenance
Traceable metadata on the origin and composition of training data....
Risk classification (of AI applications)
Assessment of AI applications by materiality and risk factors....
Data Provenance
Documentation of data origin, collection, and transformations for AI systems....
privacy risk
Likelihood × impact of privacy harms....
Risk assessment
Procedure to identify and mitigate social, ethical and legal risks....