Fairness and privacy in voice biometrics: A study of gender influences using wav2vec 2.0