Overview
EN ISO/IEC 8183:2024 / ISO/IEC 8183:2023 defines a standardized data life cycle framework for artificial intelligence (AI) systems. Adopted by CEN as EN ISO/IEC 8183:2024, the standard maps the stages data pass through from initial idea conception to final system and data decommissioning. It is technology‑agnostic and does not mandate specific tools, services or platforms. The framework is intended to improve AI data governance, data quality, security and system utility across organizations of all sizes.
Key topics and technical requirements
- Ten life cycle stages are identified and described:
- Idea conception, Business requirements, Data planning, Data acquisition, Data preparation, Building a model, System deployment, System operation, Data decommissioning, System decommissioning.
- Life cycle processes: actions and processes appropriate to each stage (planning, acquisition, preparation, verification/validation, maintenance, decommissioning).
- Verification vs validation: distinction between internal model verification/validation and whole‑system validation during operation.
- Data decommissioning vs system decommissioning: Stage 9 focuses on data-specific actions (secure deletion, archiving, repurposing) while Stage 10 covers disposal of the system regardless of data outcome.
- Terminology and references: aligns with ISO/IEC 22989 (AI concepts and terminology) and references other related work (e.g., ISO/IEC 23053, ISO/IEC 5212 under preparation).
- Applicability and constraints: applicable to all organizations using data in AI development and operation; highlights patent considerations and national adoption rules under CEN/CENELEC.
Practical applications
- Establishing repeatable AI data governance and lifecycle management processes.
- Designing compliant workflows for data planning, acquisition and preparation to reduce bias and improve data quality.
- Guiding teams on model building, verification/validation and safe system deployment/operation.
- Defining secure data decommissioning and archival procedures for privacy and compliance (e.g., PII handling, DPIA considerations).
- Informing procurement, auditing and risk management for AI projects.
Who should use this standard
- AI/ML engineers, data engineers and data scientists
- IT architects and system integrators
- Compliance, privacy and security teams
- Product managers and risk officers overseeing AI systems
- Standards bodies and organizations implementing AI governance
Related standards
Keywords: ISO/IEC 8183, data life cycle, AI data governance, artificial intelligence data framework, data decommissioning, model verification, AI system lifecycle.
