A company uses a machine learning model to analyze security camera footage for theft detection. The model has been found to incorrectly flag individuals from a particular ethnic group at a significantly higher rate than others. What category of bias is influencing the model's predictions?