ASR Auditing Flaws Masked Harm to Aphasia Speakers, Study Finds

A new research paper published on arXiv identifies critical flaws in current auditing practices for Automatic Speech Recognition (ASR) systems. The study highlights how standard methods can obscure performance disparities, particularly for individuals with speech disorders like aphasia. The authors propose a more comprehensive auditing framework to address issues such as inadequate text standardization, insufficient subgroup analysis, and the over-reliance on single metrics, advocating for community-driven approaches to ensure equitable ASR performance. AI

IMPACT Highlights need for more equitable AI auditing, especially for marginalized user groups with speech impairments.

RANK_REASON The cluster contains an academic paper detailing research findings and proposing a new framework. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• Anna Seo Gyeong Choi
• aphasia
• arXiv

• paper
• safety
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →