Which of the following is an AI-specific step in incident response?

• A. Data provenance checks should be performed during data collection
• B. Incident containment
• C. Model retraining
• D. Drift detection after deployment

Answer: (D)

AI-specific incident response centers on monitoring deployed models for changes in behavior that signal problems. Drift detection after deployment targets data drift and concept drift that can degrade model performance once real-world data arrives or the environment changes. Detecting such drift allows the team to quickly investigate, contain any issues, and decide on remediation like retraining or rollback, which is a distinctly AI-focused step in an incident response playbook. Other options relate to data governance (provenance checks during data collection), general incident containment, or broader maintenance (retraining) and don’t capture the production-model risk in the same AI-specific way.