Control Details
Control Implementation
View control requirements, implementation steps, and evidence.
View control requirements, implementation steps, and evidence.
Examine training, validation, and testing datasets for possible biases to identify discriminatory patterns before they are learned by AI models, preventing bias propagation to AI system outputs, in compliance with EU AI Act Article 10(4).
EU AI Act
Article 10Frequency
Before training, after dataset updates
Category
Data GovernanceStandard
STD-AI-003Examine training, validation, and testing datasets for possible biases to identify discriminatory patterns before they are learned by AI models, preventing bias propagation to AI system outputs, in compliance with EU AI Act Article 10(4).
Control Type
preventive control - Designed to prevent issues before they occur
Review Frequency
Before training, after dataset updates
Risk Level
high - Important for compliance, requires prompt attention
Refer to the source standard for detailed implementation steps, evidence requirements, and success criteria.
AI Data Governance Standard →POL-AI-001
AI Governance Policy
View policy for full requirements mapping.
Control ID
DATA-007
Legal Basis
Article 10Source Standard
STD-AI-003Parent Policy
POL-AI-001Risk Level
High