Abstract

As artificial intelligence voice assistants become a mainstream form of human-computer interaction, the gender and occupational stereotypes embedded in voice design have emerged as one of the most heated sociolinguistic debates, particularly in the context of China’s strengthening AI ethics governance and the growing influence of domestic AI platforms. This study examines the linguistic features displayed on Chinese Doubao AI voice platform, focusing on voice labeling, role allocation and acoustic design. Using a mixed-methods approach, the research combines quantitative analysis of 152 structured questionnaires with qualitative thematic coding of semi-structured interviews with and open-ended responses from users’ perceptions. Drawing on the concepts of explicit-implicit stereotype theory, social gender role theory, and the yin-yang cultural gender paradigm, the study reveals that Doubao’s voice library contains clear gender and occupational stereotyping: female voices are commonly associated with gentle and service-oriented traits, while male voices are linked to authority and technology-related roles. Although most users recognize such stereotypes, many still accept or enjoy them because of emotional attachment and entertainment value. Moreover, superficial technological diversity and algorithmic recommendation systems together create a self-reinforcing cycle of stereotypes. This study contributes to the research on AI voice bias in Chinese local platforms and provides insights for AI ethical design, diversified voice ecosystems, and gender equality governance in technology.

This work is licensed under a Creative Commons Attribution 4.0 License.