Using voice and speech data in healthcare: a scoping review of the ethical, legal and social implications

Human voice and speech, integral to personal identity and social communication, are increasingly used as biometric and digital biomarkers in healthcare. Their collection and analysis, enabled by artificial intelligence, machine learning, and natural language processing, offer promising applications in disease detection and health monitoring. This scoping review examines the ethical, legal, and social implications (ELSIs) associated with using voice and speech data in healthcare. Following a structured search of four databases and a snowball method, 65 articles published between 2009 and 2024 were analyzed. The findings are organized into three main ELSI categories: ethical concerns include privacy breaches, challenges of informed consent, and the need for data validation and respect for vulnerable populations; social issues highlight biases, representational disparities, and risks of discrimination and data misuse; legal issues include unclear regulatory frameworks, conflicting jurisdictional mandates, and challenges in defining data ownership. The review reveals that while many ELSIs mirror those of other biomarker data, the unique properties of voice and speech require adapted frameworks for consent, data governance, and privacy protection. Technological limitations, dataset scarcity, and industry-academic divides exacerbate risks and hinder equitable development. Few studies deeply explore ELSIs in underrepresented populations, and there is a lack of robust empirical research. The review argues for a contextualist, not exceptionalist, approach to voice biomarkers, acknowledging both overlapping and unique challenges. It concludes by stressing the need for harmonized regulations, inclusive datasets, and interdisciplinary collaboration to ensure responsible, equitable integration of voice and speech technologies in healthcare.