arXiv:2606.26901v1 Announce Type: cross Abstract: Automatic Speech Recognition (ASR) is increasingly used to document clinical encounters, yet its reliability in multilingual and demographically diverse Indian healthcare context remains largely unknown. In this study, we first conduct the systematic audit of ASR performance on real-world psychiatric interview data spanning Kannada, Hindi and Indian English, comparing eight state-of-the-art models including IndicWhisper, WhisperLargeV3, Sarvam, GoogleS2T, Gemma3n, OmniLingual, Vaani, and Gemini. Our results reveal substantial variability across
The increasing deployment of AI in sensitive applications like healthcare, coupled with the rising availability of multilingual models, necessitates immediate assessment of their reliability and biases in diverse linguistic and demographic contexts.
This study highlights critical performance disparities in clinical AI for non-Western languages, directly impacting healthcare equity and the safe deployment of AI systems globally, especially as AI adoption accelerates in emerging economies.
The understanding of current state-of-the-art ASR model limitations in multilingual clinical settings for Indian languages changes, emphasizing the need for localized development and auditing.
- · Local AI developers for Indian languages
- · Healthcare providers in India
- · Patients in diverse linguistic regions
- · Clinical AI auditing firms
- · Generic, unadapted global ASR models
- · Healthcare systems relying on un-audited AI
- · Companies neglecting local language data
Increased focus and investment in developing region-specific and language-specific AI models for healthcare applications.
Heightened regulatory scrutiny and the development of localized ethical AI guidelines for clinical use in multilingual environments.
A potential 'fragmentation' of the global AI market, with a strong emphasis on sovereign AI solutions tailored to specific linguistic and cultural contexts, especially in critical sectors like healthcare.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI