Measure of potential hazards that may arise from the use of an AI system in a specific application context. The term is often used in a similar way to risk, with criticality focusing more on an assessment of the overall system.

Note: see „Kritikalität, Kritikalitätspyramide, risikoadaptierter Regulierungsansatz“ in „Gutachten der Datenethikkommission“ 10/2019¹ and „Abschlussbericht Enquete-Kommission KI“ 10/2020².

See also:

• risk