An AI-based marketing analytics tool is trained on data that is five years old.

Model drift (also called concept drift) occurs when the data used for training no longer reflects the current environment. In marketing, consumer behaviors, trends, and demographics change rapidly; therefore, a model trained on five-year-old data will suffer from significant drift. The AAIA™ manual explains that this results in the model making predictions based on obsolete patterns, leading to poor accuracy and potentially biased or irrelevant marketing recommendations. Options A, B, and C refer to security attacks, whereas " Model drift " is the natural operational degradation that occurs when AI systems are not regularly updated with fresh, representative data.