Assess whether a shadow system must be decommissioned this quarter (57d266)

Situation

A bank preparing for a model-risk exam cannot treat a new use case bolted onto a model approved for a narrower purpose as color commentary on vendor model card with missing evaluation slices. Vendor-diligence reviewer for AI tools must close a shadow system must from that extract under AI Governance / Bias and Training Data.

Decision

Vendor-diligence reviewer for AI tools in a bank preparing for a model-risk exam must choose Policy or governance breach / Model defect / Dual failure / Hold for the missing fact using vendor model card with missing evaluation slices after a new use case bolted onto a model approved for a narrower purpose.

Hypotheses to test

• Authorize Policy or governance breach now; vendor model card with missing evaluation slices already has the discriminator after a new use case bolted onto a model approved for a narrower purpose.
• Keep Model defect in force until vendor model card with missing evaluation slices is completed after a new use case bolted onto a model approved for a narrower purpose for vendor-diligence reviewer for AI tools.
• Treat vendor model card with missing evaluation slices as Dual failure because both readings appear after a new use case bolted onto a model approved for a narrower purpose.
• Refuse a AI Governance close: vendor-diligence reviewer for AI tools does not have the page a shadow system must turns on in vendor model card with missing evaluation slices.

Analysis required

• Check intended purpose and inventory status against EU AI Act / exam-readiness language after a new use case bolted onto a model approved for a narrower purpose.
• Map the approved-use case to the system a shadow system must would bind.
• Check intended purpose and inventory status against EU AI Act / exam-readiness language after a new use case bolted onto a model approved for a narrower purpose.
• For this AI Governance Bias and Training Data file, read vendor model card with missing evaluation slices against a new use case bolted onto a model approved for a narrower purpose and write the one fact that would move a shadow system must for vendor-diligence reviewer for AI tools.

Recommendation

Choose Policy or governance breach / Model defect / Dual failure / Hold for the missing fact on this AI Governance / Bias and Training Data packet (vendor model card with missing evaluation slices after a new use case bolted onto a model approved for a narrower purpose). Lead with the AI Governance option vendor model card with missing evaluation slices can support after a new use case bolted onto a model approved for a narrower purpose, then the two facts that force it, then the Monday action for vendor-diligence reviewer for AI tools in a bank preparing for a model-risk exam.

Command returns

• Bottom-line AI Governance option on a shadow system must, then the evidence in vendor model card with missing evaluation slices, then the action for vendor-diligence reviewer for AI tools
• Hypothesis scorecard against vendor model card with missing evaluation slices: supported / rejected / untestable
• Regulatory or exam hook Bias and Training Data would cite
• Bias and Training Data finding in vendor model card with missing evaluation slices that a second reviewer can re-perform