A missing deployment and monitoring policy: Which lifecycle-stage policy gap most directly explains this
Undetected accuracy degradation after a successful launch signals a missing deployment-and-monitoring policy for ongoing performance tracking.
The question
An AI content-moderation system performed well at launch, but eighteen months later its accuracy on new slang has quietly degraded and no one noticed for months. The company had strong pre-launch testing policies. Which lifecycle-stage policy gap most directly explains this failure?
Preparing for AIGP? Take the free 5-min readiness quiz →
- A missing training and testing policy, since the model was evidently never validated against representative data before it was approved for release.Contradicted by the facts, which state pre-launch testing was strong; the failure emerged only after deployment.
- A missing deployment and monitoring policy, since no process tracked post-launch performance drift and triggered a timely review of the live model. ✓Correct. Undetected degradation after launch points to an absent deployment-and-monitoring policy for ongoing performance tracking.
- A missing use case assessment policy, since the moderation problem itself was apparently never scoped or justified before development work began.Plausible, but the system launched successfully, indicating the use case was scoped; the gap is post-launch, not at intake.
- A missing data acquisition and use policy, since the original training corpus was seemingly gathered without governing its sourcing and permitted uses.Almost right that data governance matters, but sourcing controls would not detect gradual accuracy drift on new slang after release.
The trap
Assuming any accuracy problem traces back to pre-launch testing, even when the failure surfaces only after deployment. How to remember it
Undetected accuracy degradation after a successful launch signals a missing deployment-and-monitoring policy for ongoing performance tracking.
How many of these would you get right?
One of 1581 AIGP questions on Certsqill. Take a free five-minute check and see your score per domain — not one number, but which section to open tonight.
Test your AIGP readiness — freeMore Understanding the Foundations of AI Governance questions
- Monitoring the model's real-world impact on its customers: From a governance perspective, which responsibility →
- Intellectual property and data governance policies: Which existing policy area most urgently needs updating →
- Use procurement assessments and contract terms setting: Which measure most directly manages this third-party →
- All 337 Understanding the Foundations of AI Governance questions →
Part of the Certsqill AIGP question bank · Understanding the Foundations of AI Governance ·
Every answer, right and wrong, comes with its own explanation.
