A growing set of international standards now defines what "responsible AI" looks like in practice. Understanding how they fit together — and where AI Safety Council certifications sit relative to them — is essential for anyone governing AI in a safety-critical environment.
The core standards landscape
ISO/IEC 42001 is the first international standard for an AI Management System (AIMS), giving organisations a structured, auditable way to govern how AI is developed, deployed and monitored. ISO/IEC 23894 provides AI-specific risk management guidance, extending the general ISO 31000 risk framework to AI's particular failure modes. NIST's AI Risk Management Framework offers a widely-adopted (if voluntary) US counterpart, organised around Govern, Map, Measure and Manage functions. Alongside these sit sector and regional instruments such as the EU AI Act, which is fast becoming a de facto global baseline for high-risk AI use, including AI used in employment and workplace settings.
Where certification fits
The AI Safety Council is not a standards body — it is a certification authority. Our role is to independently verify that individual professionals have the competence to implement, assess or investigate AI systems against these standards. An organisation can adopt ISO/IEC 42001; certifying that the people running the management system actually know what they're doing is a separate, and often overlooked, requirement.
How the three certifications map
AISP focuses on safe implementation and operation of AI systems in safety-critical environments. AIRP centres on risk identification, controls and assurance, aligning closely with ISO/IEC 23894 and NIST's Measure/Manage functions. AIIP validates the ability to investigate AI-related incidents and failures once something has gone wrong — a competence increasingly expected under both ISO/IEC 42001's continual-improvement requirements and emerging incident-reporting obligations.
See the full certification programmes for syllabus detail, or read how these standards are playing out in practice on our ISO/IEC 42001 and EU AI Act coverage.