Standard / soft law
AI systems should undergo a documented AI system impact assessment (ISO/IEC 42001)
In our own words: ISO/IEC 42001:2023 distinctively requires an AI system impact assessment — assessing potential consequences of an AI system for individuals, groups of individuals, and society — carried out and maintained within the AI management system (Clause 6 planning + the Annex A impact-assessment control area), rather than only an organizational risk assessment. Detect an AI deployment with no documented impact assessment covering individual, group, and societal effects. [TIER C: own-words summary of PUBLIC scope; licensed normative text not stored.]
iso-42001.ai-system-impact-assessment · ISO/IEC 42001:2023, Clause 6 + Annex A (AI system impact assessment) · official source · jurisdictions: INTL
Standard / soft law
Organizations developing, providing, or using AI should operate a certifiable AI management system (ISO/IEC 42001)
ISO/IEC 42001:2023 is the first certifiable international AI management system (AIMS) standard. In our own words: its auditable requirements (Clauses 4-10, Annex-SL harmonized structure) call for establishing context and AIMS scope, top-management commitment via an AI policy with assigned roles and authorities, AI-specific risk assessment and objectives, resources/competence/documented information, lifecycle operation controls, performance evaluation via internal audit and management review, and continual improvement. Annex A supplies a reference control set (38 controls across nine areas incl. AI policy, impact assessment, lifecycle, data governance). Applies across the AI value chain — developers, providers, and users. Detect an organization developing/providing/using AI with no documented AI management system artifacts. [TIER C: own-words summary of the standard's PUBLIC scope; licensed normative text not stored — purchase from ISO to read it.]
iso-42001.aims-management-system · ISO/IEC 42001:2023, Clauses 4-10 (AIMS requirements) · official source · jurisdictions: INTL