Can deep learning models be biased in their predictions?

Deep learning models can be biased in their predictions if the training data consist of biased information. For example, if a deep learning model used for screening job applicants has been trained with a dataset consisting primarily of white male applicants, it will consistently favor this specific population over others.