Which statement best describes model drift and how it is managed in production AI systems?

• A. Drift is a routine software update that doesn't affect performance
• B. Drift refers to hardware compatibility issues
• C. Drift is caused by data leakage
• D. Drift is when the model's performance declines due to changing data; managed via monitoring, retraining, and governance

Answer: (D)