“Labor, Power, and Belonging: The Work of Voice in the Age of AI Reproduction”. Almeda, S., Netzorg, R., Li, I., Tam, E., Ma, S., Wei, B. ACM Conference on Fairness, Accountability, and Transparency (FAccT) 2025.

Developments in Speech AI have facilitated a rise in text-to-speech (TTS) generation and voice-cloning systems that progressively require less data to synthesize higher-quality speech. As voice actors adapt to the impacts of voice-cloning and speech data collection on their industry, the researchers and developers driving these changes are removed from this affected context. Towards an informed perspective that centers a community directly impacted by Speech AI development and deployment, we interview 15 professional voice artists on their perceptions and responses to the emergence of Speech AI in their work. We identify how access to community, prior experiences negotiating creative labor rights, and existing exploitative power structures influence artists’ expectations and agency in the face of AI-driven disruption. We call for technologists to take a community-led approach, identifying opportunities to shift Speech AI towards creative empowerment, and to look beyond technology-centered solutions towards shaping a more equitable future for voice work.