HomeAI GovernanceAI Risk Assessment

🧭 AI risk assessment — responsible AI risk register

Build a responsible-AI risk register across 8 risk dimensions, informed by the NIST AI Risk Management Framework and ISO/IEC 23894 risk-management guidance. Start from a 14-item starter risk library, score each risk by likelihood × impact, record mitigations & owners, and export your register. Word + CSV.

Organization & system
8 risk dimensions

Every risk in the register below is tagged to one of these dimensions, drawing on common AI risk taxonomies such as the NIST AI Risk Management Framework (AI RMF 1.0) and ISO/IEC 23894 guidance on AI risk management.

Starter risk library (14 common AI risks)

Check any risks that apply to your system to add them to the register below, pre-filled with a description and default score — edit everything after adding.

Your risk register

Score each risk: Likelihood 1 (rare) – 5 (almost certain) × Impact 1 (negligible) – 5 (severe) = risk score. Bands: 1–4 Low, 5–9 Medium, 10–14 High, 15–25 Critical.

Register summary

⚠️ This tool is a self-assessment aid informed by the NIST AI Risk Management Framework (AI RMF 1.0) and ISO/IEC 23894 guidance on AI risk management. It is not a certification, not a regulatory filing, and not a substitute for a qualified risk, legal or AI-safety review. Likelihood × impact scores are a prioritization aid — treat any risk with severe potential impact as a priority regardless of its numeric score.