Data leakage
Also: leakage, target leakage
Information that would not be available at prediction time finding its way into training, so offline results look excellent and production results do not.
What people get wrong
The usual culprit is not the label itself but something downstream of it — a field filled in after the outcome, or a normalisation computed over the whole dataset before splitting.
A definition is not an understanding. This is taught properly in Regularisation and validation, hands on, part of Overfitting & regularisation.
Learn this properly