METM26 presentation

AI-mediated speech-to-text usage in multilingual healthcare encounters

Amy Dara Hochberg, Barcelona, Spain

This talk examines how AI-mediated, real-time speech-to-text technologies are reshaping multilingual clinical communication and redefining the role of language professionals. These systems transcribe spoken interaction instantly and display it as written text during consultations. Unlike conventional transcription software or auto-generated captions used after the fact, they operate live within the clinical encounter.

I will explain how and why these systems are currently used in healthcare contexts, e.g., to support clinicians’ note-taking, to provide on-screen text for patients, or to feed machine translation tools that render speech into another language in real time. Examples will be drawn from interactions involving English, Spanish, and Catalan, and will include reference to tools such as Google Live Transcribe and Microsoft Translator.

Crucially, these technologies do not eliminate the need for language professionals. Instead, they depend on multilingual language professionals to monitor output, correct errors, manage terminology, ensure pragmatic appropriateness, and intervene when automated processes misrepresent meaning. In some models, the language professional works alongside the technology as a linguistic mediator, editor, and risk manager during the live interaction.

Drawing on emerging practice and research, I will examine how authority, authenticity, and patient voice are negotiated when speech becomes data in real time. I argue that language professionals play a central role in safeguarding accuracy, nuance, and ethical standards in technology-mediated consultations, ensuring that patients’ voices are faithfully represented.

This session is aimed at medical linguists, translators, interpreters, and others interested in AI-mediated accessibility. It will provide practical guidance that participants can apply in their work, including ways to educate clients and to advocate for greater accessibility for deaf and hard-of-hearing people and those who are not proficient in the local language. It will also help mitigate potential harms associated with current AI-driven technologies.

About the presenter

Amy Dara Hochberg, PhD, is a multilingual health communication researcher, academic and medical translator, and EFL teacher whose lifelong profound deafness and lived experience fuel her work on language access in healthcare.