This Research Topic has been developed in collaboration with
The Bridge2AI-Voice Consortium, which is part of the Bridge2AI Program, funded by the NIH Common Fund. You can read and download their 2024 annual symposium abstracts
here and the annual symposium proceedings
here.
Voice and speech biomarkers are rapidly advancing, driven by significant breakthroughs in voice AI technology. The non-invasive nature of voice data collection makes it particularly valuable for diagnostics in diverse and often resource-limited settings or simply to overcome complex or costly clinical processes. Despite its promising applications in healthcare diagnostics and monitoring, the development and deployment of vocal/speech biomarkers and voice AI present complex challenges that span multiple disciplines. Effective collection, sharing, and analytics of voice data, along with the development, evaluation, monitoring, and integration of voice AI into healthcare systems, require not only technical innovation but also rigorous consideration of clinical, scientific, ethical, legal, societal, and implementation aspects. Addressing these challenges comprehensively is essential to ensure the development and usage of vocal/speech biomarkers and voice AI solutions that are ethically sourced, clinically valid, socially responsible, inclusive, and universally accessible.
This Research Topic is dedicated to addressing a broad spectrum of challenges associated with the development and integration of voice AI into healthcare. Our focus spans throughout the AI lifecycle and over several critical domains: clinical (enhancing the reliability of voice AI diagnostics and prediction), scientific (linking voice features with health conditions), technical (improving validation processes and AI robustness), and implementation (trustworthy and seamlessly integrating these technologies into existing healthcare workflows). Additionally, this topic tackles the challenges of equity, diversity, inclusion, and accessibility by promoting inclusive, culturally sensitive, and unbiased systems. We also explore ethical issues, regulatory complexities, data privacy protection, legal compliance across jurisdictions, societal trust, and sustainable commercial strategies for ethically translating research into significant patient outcomes. Through fostering comprehensive and interdisciplinary dialogue, this collection aims to advance voice AI in a way that is effective, equitable, and adaptable to the diverse needs of global healthcare systems.
This Research Topic invites manuscripts that delve into the multidimensional challenges of developing and implementing voice AI within healthcare settings. We welcome all
article types accepted by Frontiers in Digital Health that provide innovative solutions and discussions on these topics, contributing to the responsible and effective use of voice AI in global healthcare. Submissions that emphasize international and multicultural considerations are particularly encouraged to ensure the broad applicability and relevance of voice AI technologies.
We seek contributions that address specific themes including:
• What constitutes ethical sourced, clinically valid, socially responsible, inclusive, and accessible Voice AI;
• EDI considerations in participant recruitment, voice data collection, and algorithm training;
• Effective and scalable methods for clinical validation;
• Strategies for integrating voice AI into routine health workflows;
• Development of sophisticated algorithms capable of interpreting complex voice data and leveraging digital biomarkers;
• Technical improvements for real-time processing and accommodating diverse voice and speech features;
• Potential misuses and dual usages of voice data and voice AI, such as unauthorized re-identification (of people or speech), unapproved voice cloning, or unconsented environmental monitoring;
• Ethical frameworks and governance mechanisms to guide voice data usage and AI decision-making;
• Legal standards and ensuring compliance across different healthcare jurisdictions;
• Commercial strategies that emphasize scalability, sustainability, affordability and accessibility;
• Equitable dissemination of voice AI solutions in healthcare;
• International considerations in data collection, sharing, analysis, and usage addressing the historical divide between the Global North and South and catering to the differentiated needs of diverse cultural, ethnic, linguistic, and socio-economic communities.
Conflict of interest: Topic Editors Maria Powell and Jean-Christophe Bélisle-Pipon both receive research funding from
NIH for research into voice AI.