ISO/IEC 22989:2022 — Concepts et terminologie de l'IA

Technical explanation

The standard provides a shared vocabulary for AI, including concepts related to systems, models, data, learning, capabilities, stakeholders, and lifecycle activities. Consistent terminology supports communication across technical, governance, procurement, assurance, and regulatory work.

Business relevance

A controlled vocabulary reduces ambiguity in policies, contracts, inventories, risk assessments, and system documentation, particularly when multidisciplinary teams use the same words differently.

Implementation example

An organisation aligns its AI policy, system inventory, supplier questionnaire, and impact-assessment template with ISO/IEC 22989 terminology to prevent inconsistent classification.

Limitations and common misconceptions

Terminology standards improve consistency but do not specify controls, certify systems, or settle every emerging usage. New concepts such as generative AI may be addressed through later amendments or complementary standards.

Sujets

Sources