Signal Paper Reframes Speech-AI Failures for Indigenous Languages as Linguistic Policy, Not Just Technical Error

Summary

A preprint by Jay L. Cunningham, Mark Atta Mensah, Richard Martinez, Joao Vieira da Silva Neto, and colleagues focuses on automatic speech recognition (ASR) and voice interfaces that shape access to public services, healthcare, and education. The authors argue that persistent recognition failures for low-resource, Indigenous, and non-standard language varieties are not merely technical errors but function as implicit linguistic policies that reproduce colonial language hierarchies. Drawing on theories of linguistic capital, the paper proposes a framework for building cross-culturally competent speech AI systems. The work situates speech-AI accessibility as a matter of language justice rather than purely a benchmark performance gap.