Overview
ISO/IEC 5259-5:2025 - Artificial intelligence - Data quality for analytics and machine learning (ML) - Part 5: Data quality governance framework - defines a governance framework to help organizational governing bodies direct and oversee data quality measures, controls and processes across the data life cycle (DLC) for analytics and ML. The standard is applicable to any analytics or ML use and is intended to be used alongside other parts of the ISO/IEC 5259 series. It does not prescribe specific management or process requirements (those are covered in ISO/IEC 5259-3 and -4).
Key topics
- Data quality governance in analytics and ML: Establishing visibility and strategic direction so governing bodies understand how data quality affects ML outputs and automated decision-making.
- DQ guiding principles and policies: High-level principles and organizational strategies to align data quality goals with business objectives.
- Business planning and accountabilities: Roles and responsibilities for governing bodies, management, data owners, data stewards and operational teams across the management and operational layers.
- DQ risk management: Identification, oversight and mitigation of data-related risks that can degrade analytics and ML model performance.
- Management processes and controls: Framework elements for implementing, monitoring and adjusting data quality practices throughout the DLC in line with ISO/IEC 5259-1.
- Data and dataset quality characteristics: Attention to characteristics such as accessibility, auditability, identifiability, currentness, accuracy, balance, diversity, representativeness, timeliness and generalizability (see ISO/IEC 5259-2 for definitions).
Applications
- Boards, audit committees and governing bodies establishing strategic oversight for AI/ML deployments.
- Executive managers (CEOs, CDOs, CIOs) and risk/compliance teams implementing enterprise-level data governance and controls for ML systems.
- Data governance, data stewardship and ML operations teams designing policies, monitoring processes and accountability structures across the DLC.
- Organizations integrating third‑party or sensor data where ambiguous ownership and lineage can affect model outcomes - the framework helps clarify responsibilities and oversight.
- Use cases include regulated industries, enterprise AI programs, model risk management, and any analytics projects where data quality impacts automated decisions.
Related standards
By adopting ISO/IEC 5259-5:2025, organizations can create a structured, auditable data quality governance framework that reduces ML risks, improves model reliability, and aligns data practices with corporate oversight and compliance objectives.