GenAI Citizen Adaptors for Multilingual Health Communication

By Danielle H. Heinrichs, Michael Camit, and Jack Tsao

About

Despite multilingual health communication being a fundamental human right, the requirement for costly professional translation services creates barriers for non-native language-speaking populations. This study explores how 22 citizen adaptors (non-professional student translators) in Australia and Hong Kong collaborated with generative AI and a health NGO to adapt neurofibromatosis health messages from English into Chinese Mandarin, Cantonese, Indonesian, and Russian. Drawing on Haraway’s concept of “response-ability,” the research challenges traditional translation hierarchies by examining how citizen adaptors developed capabilities to respond through reciprocal teaching with GenAI, biliteracy skills beyond mere bilingualism, and affective connections to their audiences. While the study revealed challenges, including difficulties with terminology adaptation and time constraints that questioned assumptions about AI-aided translation being faster, it demonstrates the potential for non-professional, AI-assisted adaptation to democratise multilingual health communication when supported by appropriate training and policy frameworks.

Key Findings and Contributions

  1. Citizen adaptors developed strategies for teaching and providing feedback to generative AI systems, creating iterative improvement loops that enhanced translation accuracy through human-AI collaboration.

  2. The study identified that written language competency (biliteracy) and AI literacy were more critical than native speaker status, challenging traditional assumptions about who can perform health translations.

  3. Emotional connections to target audiences improved adaptation quality, contradicting professional translation norms that emphasise detachment and objectivity.

  4. Despite GenAI assistance, quality adaptations required significant time investment, revealing tensions between the promise of rapid AI translation and the reality of careful, culturally-responsive adaptation work.

Implications for Policy and Practice

  1. Organisations should develop classification systems that match content sensitivity with translator expertise levels, allowing citizen adaptors to handle appropriate materials while reserving high-risk content for professionals.

  2. National/ local education systems need to integrate biliteracy training for discipline-specific terminology, and AI literacy from early on to build a pipeline of capable citizen adaptors.

  3. Creating multilingual glossaries using approaches like Minimal Languages for common health emergencies (pandemics, heatwaves) could significantly reduce adaptation time and costs during crises.

  4. Current policies mandating only professional translators need revision to formally recognise and regulate the role of trained citizen adaptors, particularly for time-sensitive public health communications where professional services are unavailable.

Publication

H. Heinrichs, D., Camit, M., & Tsao, J. (2025). Response-able, collaborative adaptations of multilingual health messaging: a case study from Australia and Hong Kong, SAR. Critical Public Health, 35(1). https://doi.org/10.1080/09581596.2025.2555211