August 9, 2025

Voice for All: Evaluating the Accuracy and Equity of Automatic Speech Recognition Systems in Transcribing Patient Communications in Home Healthcare.

View Full Paper

Ask AI

Bookmark

Automatic speech recognition systems show reduced transcription accuracy for Black patients compared to White patients.
The word error rate for AWS General was 39%, the highest among the evaluated ASR systems.
Different linguistic domains, such as Affect and Social, posed significant challenges for ASR accuracy.
Diverse training datasets are necessary to enhance equity in automatic speech recognition performance within healthcare.

Discussion

No takes yet. Share an insight, caveat, or question.

Evaluation reveals transcription accuracy disparities in automatic speech recognition systems, indicating potential biases in healthcare delivery.

Xu et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark

Evaluation reveals transcription accuracy disparities in automatic speech recognition systems, indicating potential biases in healthcare delivery.