Question 970
At a model development workspace receiving external data, an AI project leader faces this situation: A financial fraud model has inconsistent ground-truth labels and data owners will not resolve disputed cases. The source-to-feature path must be supportable and auditable. Which underlying factor accounts for the described failure?
Explanation
Correct Answer
In a model development workspace receiving external data, the underlying issue is: Data readiness approval should be withheld when required quality thresholds are materially unmet. It explains the scenario: A financial fraud model has inconsistent ground-truth labels and data owners will not resolve disputed cases. The appropriate project response is: Document the failed label thresholds and escalate a no-go recommendation to the accountable sponsor. This can be checked by: The decision record identifies the blocked criterion and the owner of required corrective action.
Incorrect Answers
Source integration requires explicit entity-matching rules rather than concatenating records blindly. The appropriate correction for the reported case is: Document the failed label thresholds and escalate a no-go recommendation to the accountable sponsor.
Missing values may be systematic and carry subgroup or process bias rather than random noise. Fixing that different mechanism would not address the current root cause: data readiness approval should be withheld when required quality thresholds are materially unmet.
Leadership communication should identify actionable data access obstacles and required decision owners. Its verification is distinct from the reported condition: A financial fraud model has inconsistent ground-truth labels and data owners will not resolve disputed cases. The source-to-feature path must be supportable and auditable.