Speech Diversity

The full range of ways human speech varies from the narrow 'typical' speech on which most speech-AI systems are trained and benchmarked. Speech diversity includes people who stutter, d/Deaf and Hard-of-Hearing speakers, people with dysarthria, aphasia, or other neurological speech conditions, age-related voice changes, second-language and accented speakers, gendered speech patterns, and racial, ethnic, and regional vernaculars and dialects. Speech AI systems that ignore speech diversity systematically exclude users from everyday and high-stakes voice interactions — voice assistants, medical transcription, job interview platforms, meeting-transcription tools, AI-mediated AAC. Addressing speech diversity requires inclusive training data, subgroup-specific evaluation metrics, and participation of diverse speakers in design and measurement.