Question 350
At fictional Meridian Civic Exchange, a Snowflake data and AI engineering team is conducting a controlled model-deployment decision with regional service policies and accountable stakeholders. The separately composed technical incident is: A scheduled summarization workload uses an administrative role because its initial proof of concept lacked fine-grained grants, allowing the same identity to read unrelated datasets. Assume only these documented facts and use safe, authorized production-change procedures; exact current Snowflake feature syntax and region or account availability require current review. Which repeatable assurance or control is most relevant to preventing this situation?
Explanation
Writer-proposed answer: B (UNVERIFIED)
Preliminary individual option explanations — NOT expert validated
Correct Answer
Maintain privilege inventories, role attestations and periodic unauthorized-access regression tests for AI service accounts. Least privilege applies to AI pipelines and scheduled services exactly as it does to interactive SQL workloads.
Incorrect Answers
This response concerns Redaction quality on complex identifiers, which is different from the stated failure. A scheduled summarization workload uses an administrative role because its initial proof of concept lacked fine-grained grants, allowing the same identity to read unrelated datasets.
Agent tool execution requiring approvals is not the cause of the reported failure: A scheduled summarization workload uses an administrative role because its initial proof of concept lacked fine-grained grants, allowing the same identity to read unrelated datasets.
Audit trail for AI-generated business decisions alone does not establish the needed evidence: Functional tests prove required summaries still succeed and unauthorized table access is denied for the scheduled service identity.