Voice-based artificial intelligence (AI) is often presented as neutral infrastructure, yet its voices, names, and interaction scripts carry gendered expectations. This article examines how feminized voice AI is governed and reproduced in Indonesia. It uses a qualitative single-case study and document and discourse analysis of an 11-document corpus (2021-2025) comprising academic studies, UNESCO and Indonesian governance texts, a corporate statement, and public reporting on Project Nirmala. Guided by technofeminism, the analysis traces design, discourse, governance, and reproduction. It finds, first, that corporate reforms such as removing a female default address interface choice without necessarily changing feminized service scripts. Second, Indonesia's 2023 AI Ethics Circular recognizes inclusion, humanity, transparency, and accountability but provides no operational criteria for gendered persona or interaction design. Project Nirmala illustrates how a young female AI persona is reproduced in local innovation. The article contributes to digital International Relations (Digital IR) by showing that AI interfaces are sites where transnational corporate norms, international ethical standards, and domestic soft law interact to normalize or contest gendered power.
Copyrights © 2026