detecting a change in performance of an ML-Module over a period of time using a labeled input dataset for the ML-Module, a target value for the ML-Module, and an output value of the ML-Module, the output value being generated using the labeled input dataset with the ML-Module, the labeled input dataset being provided individually to the ML-Module over the period of time;

detecting a change in predicted performance of the ML-Module over the period of time by a drift module, the drift module being configured to compute a single value of the predicted performance using each input dataset of a set of input datasets, the input datasets of the set of input datasets being provided individually to the drift module over the period of time;

determining a value of a first key figure, the value of the first key figure indicating a correlation between the change in performance of the ML-Module and the change in predicted performance of the ML-Module; and

providing a signal that indicates the value of the first key figure.