Deployment and monitoring: Within the AI life cycle, which stage does this control most directly address?

Continuous post-launch performance tracking and drift detection is the deployment-and-monitoring stage.

The question

Preparing for AIGP? Take the free 5-min readiness quiz →

1. Incident management, where a specific realized failure or harm is investigated, escalated and then remediated.

   Plausible but wrong: incident management reacts to a triggered event, whereas monitoring is the continuous watch that may trigger it.
2. Deployment and monitoring, where a live system's ongoing performance is tracked for drift or issues. ✓

   Correct: continuous post-release performance tracking and drift detection is the deployment-and-monitoring stage.
3. Use case assessment, where the initial suitability and rationale of an AI project are evaluated.

   Plausible but wrong: use-case assessment happens before build, not during live monitoring.
4. Data acquisition and use, where training data is sourced, validated and prepared for the model.

   Plausible but wrong: this precedes deployment and concerns input data, not live performance tracking.

How to remember it

Continuous post-launch performance tracking and drift detection is the deployment-and-monitoring stage.

More Understanding the Foundations of AI Governance questions

• Deployer, because it puts an existing AI system into use: From a governance perspective, in what role is the →
• Evaluate and update those existing policies so they: Which approach best reflects sound governance as it →
• Use procurement assessments and contract terms that impose: To govern the third-party risk this introduces, →
• All 337 Understanding the Foundations of AI Governance questions →