A usable loop for identifying, scoring and treating AI risks — with a note on where governance ends and risk work begins.
AI risk management is the ongoing process of finding what could go wrong with your AI use, deciding how much it matters, and taking action proportionate to the risk.
Common dimensions: data sensitivity, autonomy of the system, reversibility of harm, affected population, and regulatory exposure. Keep the scale simple — a three-level scale used consistently beats a complex one nobody maintains.
Record the assessment, who approved it and when. That record is what turns risk work into governance evidence. See the framework guide for how this fits a broader program, and the tool comparison if you want platforms that automate it.
Looking for tools? See our comparison of AI governance platforms.