The voice signal can identify a human being. For this reason, the European Union’s (EU) General Data Privacy Regulations (GDPR) categorizes voice recordings as personal data and subject to protection. GDPR Article 4.1 defines personal data as “all data related to a person and that allows them to be identified directly or indirectly (…) to one or several specific properties unique to their physical, physiological identity…”. The voice signal is particularly unique given that it often reveals other information such as the ethnic origin of that person or health conditions such as Parkinson’s disease. This makes voice data part of a special category subject to additional protections under the GDPR than other types of personal data. > [!Main ethical considerations for Speech Technologies:] > > - Don't record someone's voice without their explicit consent > - Don't clone someone's voice without their explicit consent > - Consider minority groups and language bias when building speech-processing applications