A Sociolinguistic Analysis of Automatic Speech Recognition Bias in Newcastle English
Summary: arXiv:2603.24549v1 Announce Type: cross
Abstract
Automatic Speech Recognition (ASR) systems are widely used in everyday communication, education, healthcare, and industry, yet their performance remains uneven across speakers, particularly when dialectal variation diverges from the mainstream accents represented in training data. This study investigates ASR bias through a sociolinguistic analysis of Newcastle English, a regional variety of North-East England that has been shown to challenge current speech recognition technologies.
Research Methodology
Using spontaneous speech from the Diachronic Electronic Corpus of Tyneside English (DECTE), we evaluate the output of a state-of-the-art commercial ASR system and conduct a fine-grained analysis of more than 3,000 transcription errors. Errors are classified by linguistic domain and examined in relation to social variables including gender, age, and socioeconomic status.
Findings
The results show that phonological variation accounts for the majority of errors, with recurrent failures linked to dialect-specific features. The key findings include:
- Vowel quality and glottalisation significantly impact transcription accuracy.
- Local vocabulary and non-standard grammatical forms contribute to misrecognition.
- Error rates vary across social groups, with higher frequencies observed for men and for speakers at the extremes of the age spectrum.
Social Patterns in ASR Errors
These findings indicate that ASR errors are not random but socially patterned. The analysis reveals how linguistic characteristics interact with social variables to influence ASR performance. Specifically:
- Men exhibited a higher rate of transcription errors compared to women.
- Older adults and younger speakers faced the most significant challenges in ASR accuracy.
Acoustic Case Study
In addition, an acoustic case study of selected vowel features demonstrates how gradient phonetic variation contributes directly to misrecognition. This highlights the need for ASR systems to adapt to the phonetic subtleties of regional dialects.
Conclusion
The study emphasizes the importance of incorporating sociolinguistic expertise into the evaluation and development of speech technologies. To create more equitable ASR systems, explicit attention must be given to dialectal variation and community-based speech data. Addressing these issues will not only improve ASR performance across diverse speakers but also promote inclusivity in technology design.