Explain model drift in marketing AI models and how you would detect and respond to it.

• A. Model drift occurs when data distributions or relationships change; detect via monitoring metrics (accuracy, calibration, lift) and retrain/update when drift exceeds thresholds.
• B. Model drift occurs when a model overfits training data; respond by deleting the model.
• C. Model drift refers to changes in code version; respond by patching software.
• D. Model drift is when data volumes increase; ignore.

Answer: (A)