Glossary

AI risk management

AI risk management is the ongoing process of identifying, analyzing, evaluating, treating, and monitoring risks arising from AI systems across their lifecycle — including technical, operational, legal, security, and societal harms relevant to the organization’s context. It operationalizes controls and residual-risk decisions inside (or alongside) an AI governance system; it does not replace governance’s allocation of authority and accountability.


Why it matters

Enterprises often stop at policy documents. Risk management requires concrete threat and failure analysis for each use case: data leakage, harmful outputs, wrongful automated decisions, agent actions outside policy, third-party model changes, and concentration risk. The NIST AI RMF organizes related functions (Govern, Map, Measure, Manage) as voluntary guidance for structuring that work.

Critical distinction from AI governance: governance defines who decides and which rules bind; risk management produces the assessments, mitigations, monitoring, and residual-risk acceptances those decision-makers use. Overlap is expected; synonym collapse is not.

Hypothetical illustration: Before enabling an agent that can issue customer credits, a team maps failure modes (wrong account, prompt injection, tool misuse), measures evaluation gaps, and manages residual risk with spend caps, human approval above a threshold, and logging. The AI governance board still owns the go/no-go authority; risk management supplies the evidence pack.

Mechanisms

  • Risk identification and mapping — use-case context, stakeholders, and harm scenarios.
  • Measurement and evaluation — tests, red-teaming, monitoring metrics, and incident signals.
  • Treatment and monitoring — controls, residual-risk acceptance, and continuous reassessment after model or process change.

Boundaries

  • AI risk management ≠ AI governance as exact synonyms.
  • AI risk management ≠ “AI risk” as an exact alias — “AI risk” is informal shorthand for hazard classes; this entry names the management cycle.
  • Completing a one-time checklist ≠ ongoing risk management across model updates and new tools.
  • Framework alignment ≠ elimination of residual risk or automatic regulatory compliance.
  • Model-provider safety features ≠ enterprise risk management for a specific workflow and data boundary.

Sources

  1. National Institute of Standards and Technology (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0). NIST AI 100-1 — voluntary framework for mapping, measuring, and managing AI risks.
  2. ISO/IEC 42001:2023 — AI management system requirements that intersect with organizational risk and control processes (standard; not a substitute for legal advice).
  3. ISO 31000:2018 — Risk management — Guidelines — general risk-management vocabulary useful as background (not AI-specific).



Further research


Continue exploring

A short path from this concept into related research and practical resources.

  1. Related conceptAI governance
  2. Deeper researchThe Agentic Operating Model Building
  3. Apply in DiscoverThe Enterprise AI Playbook Library
AI Risk Management — Definition | The Strategy Stack