Inadequate planning of performance requirements
The expected performance of the AI system should be planned adequately. Hereby, an important aspect is that chosen performance metrics are meaningful for presenting the intended functionality. Otherwise, expectations and safety requirements can be unfulfillable at later life cycle stages.
ENTITY
1 - Human
INTENT
2 - Unintentional
TIMING
1 - Pre-deployment
Risk ID
mit996
Domain lineage
7. AI System Safety, Failures, & Limitations
7.3 > Lack of capability or robustness
Mitigation strategy
1. Mandate the formal definition of Specific, Measurable, Achievable, Relevant, and Time-bound (SMART) performance, safety, and fairness metrics *prior* to system development, ensuring the chosen quantitative measures are robust proxies for the intended functionality and ethical requirements. 2. Institute comprehensive, multi-stakeholder pre-deployment risk and requirements assessments, including independent third-party validation, to verify the adequacy and alignment of the planned performance metrics with anticipated real-world operational and safety constraints. 3. Implement continuous monitoring and observability frameworks to track the system's runtime performance against the established metrics, establishing a formal governance protocol for systematic review and requirements re-baselining in the event of performance degradation or model drift.