An organization is deploying an AI system to detect anomalou…

An оrgаnizаtiоn is deplоying аn AI system to detect anomalous user activity across its enterprise. The model performs well during offline testing but generates excessive false positives in production. Investigation reveals that the training data consisted primarily of historical transaction logs and static user attributes, while the deployed system operates in a dynamic environment with rapidly changing user behavior and real-time signals. Which of the following explanations best accounts for the system’s failure from a data-centric perspective?