Voice tech and AI: Is Detecting Diseases Based on 45 s of Voice Accurate? (Henry O'Connell)

Ambient documentation is becoming normal in clinics. But the most interesting “voice” capability may not be transcription at all.In the latest episode of Faces of Digital Health, Henry O'Connell (Canary Speech) explains why voice biomarkers stalled for decades: the field analyzed words, not the neurological signal behind speech production.Canary’s approach focuses on the “primary data layer”—how the central nervous system drives respiration, vocal cord vibration, and articulation in real conversational speech. A few details that stood out: ⏱️ ~45 seconds of conversation can be enough for assessment 🎛️ 2,590 voice features analyzed every 10ms (millions of data points) 🎯 Reported accuracy: 98%+ for progressive neurological conditions (e.g., Parkinson’s/Huntington’s/Alzheimer’s), while behavioral health tends to be lower (often in the 80s) 🌍 Validation is repeated per language/culture—no “deploy and hope” model 🧭 Use cases go beyond diagnosis: screening in primary care, clinical trials outcome tracking, and even in-room aggression risk signals to help protect staff One line that captures the idea: it’s about measuring what’s present in the moment—objective signals that complement clinical judgment. Time stamps: 00:00 Introduction to Voice Biomarkers in Digital Health 01:48 Historical Context and Evolution of Voice Analysis 06:52 Innovative Approaches to Voice Data Analysis 08:54 Technical Insights into Voice Analysis 16:07 Accuracy and Efficacy of Voice Biomarkers 28:27 Challenges and Acceptance in Clinical Practice 35:04 Ethical Dilemmas in Genetic Testing 36:32 Understanding Genetic Information and Its Implications 37:58 Objective vs. Subjective Assessments in Mental Health 39:59 Proactive Care and Early Detection of Cognitive Decline 42:43 Technology in Wellness and Employee Mental Health 45:18 Data Privacy and Ethical Considerations in Health Tech 49:06 Remote Monitoring and Clinical Trials 01:00:57 Future of Health Technology and Global Expansion Youtube: Website: Newsletter:

You can listen to Voice tech and AI: Is Detecting Diseases Based on 45 s of Voice Accurate? (Henry O'Connell) online on Radio and Podcast. Open the player on this page to stream the available audio.

Voice tech and AI: Is Detecting Diseases Based on 45 s of Voice Accurate? (Henry O'Connell) is an episode from Faces of Digital Health by Tjasa Zajc.

This episode is 70:24 long.

This episode was published on Jan 22, 2026.

Yes. Use the heart button on the episode page to add it to your favorite episodes list.

Yes. This page shows related episodes from Faces of Digital Health when more episodes are available from the podcast feed.