ISO/IEC 22989
The international dictionary of AI, defining what words like 'machine learning', 'neural network', and 'explainability' mean in a technically precise and internationally agreed way.
The Simple Version
The international dictionary of AI, defining what words like 'machine learning', 'neural network', and 'explainability' mean in a technically precise and internationally agreed way.
Detailed Explanation
ISO/IEC 22989:2022 (Artificial intelligence. Concepts and terminology) provides a comprehensive taxonomy of AI concepts used across the ISO/IEC AI standards portfolio. It defines foundational terms including AI system, machine learning, deep learning, natural language processing, explainability, interpretability, and bias. The definitions align with and inform the EU AI Act Article 3 definitions and the OECD AI Principles. The standard is developed by ISO/IEC JTC 1/SC 42 and serves as the definitional anchor for the broader AI standards ecosystem.
Key Characteristics
- Provides internationally agreed definitions for core AI concepts and terminology
- Developed by ISO/IEC JTC 1/SC 42, the international AI standards committee
- Informs EU AI Act Article 3 definitions and OECD AI Principles
- Foundational reference for the ISO/IEC AI standards portfolio
Why It Matters
Legal, compliance, and governance teams can use ISO/IEC 22989 to establish precise, internationally recognised definitions in AI policy documents, contracts, and audit frameworks, reducing definitional disputes.
Real-World Analogy
Like a medical terminology standard (ICD or SNOMED), providing precise, shared definitions that enable clear communication between different stakeholders working in the same domain.
Common Misconceptions
- ISO/IEC 22989 definitions are identical to EU AI Act definitions, the Act aligns with but does not replicate the standard; legal definitions in the Act are authoritative for regulatory purposes.
- Terminology standards like 22989 are only useful for academics, precise shared vocabulary is essential for any cross-functional or cross-organisation AI governance work.