An AI safety case is a documented argument, supported by evidence, that an AI system is acceptably safe for a defined use and operating context. It is especially relevant for higher-risk sectors where assurance must be explicit, reviewable, and maintained over time.