Which of the following BEST ensures AI components are validated during disaster recovery testing?

The correct answer is D. Monitoring model performance during failover and recovery. Disaster recovery (DR) testing for AI systems must confirm that models continue to produce accurate, consistent outputs during and after a failover event. Directly monitoring model performance throughout the failover and recovery cycle is the only option that validates end-to-end

AI Security Assurance and Resilience

Question

Options

ARunning simulated data-loss scenarios by deleting test feature-store records
BDisconnecting model training clusters to test retraining workflows
CSimulating DoS attacks on AI APIs
DMonitoring model performance during failover and recovery

How the community answered

(52 responses)

A
4% (2)
B
8% (4)
C
15% (8)
D
73% (38)

Explanation

Disaster recovery (DR) testing for AI systems must confirm that models continue to produce accurate, consistent outputs during and after a failover event. Directly monitoring model performance throughout the failover and recovery cycle is the only option that validates end-to-end AI component integrity under real DR conditions. Deleting feature-store records and disconnecting training clusters are destructive actions that do not simulate realistic DR scenarios; simulating DoS attacks tests availability but not model correctness or recovery validation.

Topics

#Disaster Recovery Testing#AI Model Validation#Operational Resilience#Model Performance Monitoring

Community Discussion

No community discussion yet for this question.

Full AAISM Practice