Label Drift refers to changes in the statistical properties of target labels over time, which can degrade model performance if not detected and addressed. It is common in dynamic environments like finance and marketing.