What is a challenge associated with AI model drift?

• A. The need for automated monitoring and retraining of AI models
• B. Models naturally stabilize over time
• C. Drift reduces data privacy concerns
• D. Drift eliminates need for monitoring

Answer: (A)

Model drift happens when the data the model encounters in production changes from what it learned on during training, so its predictions lose accuracy. The challenge is keeping the model reliable over time, which requires automated monitoring to detect drift and automated retraining or updating to adapt to new patterns. Without this ongoing oversight, performance can decline silently, leading to biased or wrong results. Setting up drift detection, continuous performance monitoring, data versioning, and a robust retraining pipeline helps the model stay aligned with current data.

Choices that suggest stabilization, reduced privacy concerns, or that monitoring isn’t needed don’t fit, because drift implies the opposite—continuous change and the need for active monitoring and updates.