Assess whether a shadow system must be decommissioned this quarter (903ff8)

Situation

Policy and Oversight work in a university licensing an AI proctoring vendor now turns on a shadow system must because an internal audit finding that human review logs are empty put training-data provenance questionnaire in play. Model-risk officer should say what training-data provenance questionnaire proves.

Decision

Model-risk officer in a university licensing an AI proctoring vendor must choose Policy or governance breach / Model defect / Dual failure / Hold for the missing fact using training-data provenance questionnaire after an internal audit finding that human review logs are empty.

Hypotheses to test

• An internal audit finding that human review logs are empty is noise around an already-controlled Policy and Oversight process in a university licensing an AI proctoring vendor, given training-data provenance questionnaire.
• An internal audit finding that human review logs are empty is the event in training-data provenance questionnaire that forces Policy or governance breach for model-risk officer under AI Governance.
• Training-data provenance questionnaire shows a one-file miss after an internal audit finding that human review logs are empty, not a Policy and Oversight program failure.
• Training-data provenance questionnaire cannot decide a shadow system must yet after an internal audit finding that human review logs are empty; hold is the only AI Governance close a university licensing an AI proctoring vendor can defend.

Analysis required

• Reproduce the incident row in training-data provenance questionnaire and say whether it ever touched production data.
• Split policy-or-governance failure from a model defect using prompts, outputs, and human edits in training-data provenance questionnaire.
• Reproduce the incident row in training-data provenance questionnaire and say whether it ever touched production data.
• For this AI Governance Policy and Oversight file, read training-data provenance questionnaire against an internal audit finding that human review logs are empty and write the one fact that would move a shadow system must for model-risk officer.

Recommendation

Choose Policy or governance breach / Model defect / Dual failure / Hold for the missing fact on this AI Governance / Policy and Oversight packet (training-data provenance questionnaire after an internal audit finding that human review logs are empty). The follow-on Policy and Oversight action is what model-risk officer does next: implement the option, assign an owner, and log the missing fact.