A county AI program names data stewards for warehouses but assigns none to model feature stores. What gap should audit highlight?
Select an answer to reveal the explanation.
Short Explanation
Warehouses have stewards; feature stores are orphaned — that is the gap. AI-specific assets need named owners too. If nobody owns the features, nobody owns the risk.
Full Explanation
A county AI program that names data stewards for warehouses but assigns none to model feature stores leaves AI-specific assets outside formal accountability. Audit should highlight extending data stewardship coverage to AI-specific assets such as feature stores because those stores define the features models consume, drive quality and access decisions, and concentrate high-value derived data that traditional warehouse stewardship may never touch. Named stewardship for feature stores clarifies ownership of allowable use, quality SLAs, access reviews, and incident response when feature pipelines fail or leak. Including feature definitions in warehouse quality reviews under current steward assignments fails by concept: warehouse stewards are not assigned to feature stores, and reviewing feature definitions within existing warehouse reviews does not create named stewardship or accountability for the feature store itself. Mapping feature store lineage into warehouse catalog records for end-to-end traceability fails by concept: documenting lineage in a warehouse catalog does not assign an accountable steward to the feature store, and traceability alone does not govern allowable use, quality, or access for AI-specific assets. Enforcing warehouse access controls on feature store reads for consistent data protection fails by concept: access controls protect reads but do not name a steward for the feature store, and protection without ownership still leaves feature quality, retention, and incident response without an accountable owner. Exam caveat: AAIA privacy and data-governance items treat feature stores, embeddings stores, and similar AI assets as in-scope stewardship targets—warehouse coverage alone is incomplete. Operational check: inventory AI data assets including feature stores, map each to a named steward and backup, verify stewardship RACI covers access, quality, retention, and allowable use, and confirm stewardship reviews appear in the county AI assurance plan.