B2AI-Voice is generating a robust large-scale, voice dataset to fuel discovery of voice as a biomarker of health. Through collaborative research, development of best practices, education, and events we advance the science and application of voice AI for better patient outcomes.
The Bridge2AI-Voice Consortium is dedicated to advancing voice as a powerful biomarker for health by building diverse, ethically sourced datasets and predictive AI models. Our mission is to fuel discovery of voice technologies that enable earlier detection, better treatment, and better patient outcomes through accessible and affordable care.
Voice biomarkers have the vast potential within low-cost, accessible tools for early disease detection and monitoring—potentially reaching patients anywhere in the world through simple phone-based applications.
The Bridge2AI-Voice consortium was funded by the NIH Common Fund with the mission to pioneer the most robust and ethically sourced voice dataset linked to health information, pioneering the use of voice as a biomarker of health.
In 2023, with the foundations built for rigorous team science, we began building best practices, validated multi-disorder protocols, and created software applications specific to voice data collection for health.
In 2024, our consortium expanded outreach, community engagement and built an international community interested in voice biomarkers.
The Voice AI Symposium became the premier event for researchers, clinicians, scientists, startups and industry partners interested in voice AI for healthcare.
In 2025, we are accelerating data generation across five disease cohorts, as well as controls. We are conducting regular dataset releases and releasing XAI tools and resources to facilitate use of our dataset.
Laryngeal cancers, Vocal fold paralysis, Benign laryngeal lesions
Alzheimer’s, Parkinson’s, ALS
Depression, Anxiety, Schizophrenia, Bipolar Disorders,
Pneumonia, COPD, Asthma
Autism, Speech Delay
Our dataset is designed to facilitate rigorous scientific research in voice-powered health. It includes thousands of high quality voice recordings, clinically validated by physicians and speech pathologists. Researchers can gain access to data collected using our multi-disorder protocol which features numerous acoustic tasks, 22 validated questionnaires, and medical information across 5 disease cohorts and controls. Every recording was collected under rigorous ethical oversight. The B2AI-Voice dataset is one of the most comprehensive and reliable resources available for advancing discovery of voice biomarkers.
Beyond data, we provide the resources you need to do voice AI research with confidence. Explore standardized protocols, a detailed data dictionary, and explainable AI (XAI) tools designed to make results transparent. Access best practices, ethical guidelines, and publications created by leading experts—so whether you’re new to voice AI or deep in the field, you have trusted guidance to design, interpret, and advance your studies.
Be part of the growing global community advancing voice biomarkers in health. Attend our annual Voice AI Symposium—the premier international event for discovery and collaboration. Connect with peers through open calls for science, explore publication opportunities with Frontiers in Digital Health, and stay informed about upcoming trainings, hackathons, and new research opportunities. Join our listserv and follow us on social media to engage year-round and never miss a chance to collaborate and contribute.
Join our mission to shape the future of voice in health. Collaborate, contribute, and explore today.