VASA-1: Pioneering Real-Time, Lifelike Audio-Driven Talking Faces and Its Ethical Implications