Skip to main content
Home/Risks/IBM2025/Incorrect risk testing

Incorrect risk testing

Sub-category
Risk Domain

Inadequate regulatory frameworks and oversight mechanisms that fail to keep pace with AI development, leading to ineffective governance and the inability to manage AI risks appropriately.

"A metric selected to measure or track a risk is incorrectly selected, incompletely measuring the risk, or measuring the wrong risk for the given context."

Supporting Evidence (1)

1.
"If the metrics do not measure the risk as intended, then the understanding of that risk will be incorrect and mitigations might not be applied. If the model’s output is consequential, this might result in societal, reputational, or financial harm."

Other risks from IBM2025 (63)