Designing Beyond Language: Sociotechnical Barriers in AI Health Technologies for Limited English Proficiency
AI in healthcare must do more than translate
Millions of patients in the U.S. who prefer Spanish or other languages face barriers that go far beyond words. This study asked 14 patient navigators how AI tools could help—or harm—care for Spanish-speaking patients with limited English proficiency (LEP).
- Potential helps: on-demand translation, visit prep, education, and smoother care coordination.
- Real risks: cultural misunderstandings, privacy worries, misinformation, and replacing human rapport.
- Hidden hurdles: unstable internet/phone access, low digital literacy, and clinic rules that can block new tools.
Design takeaways: build trust first (clear privacy, human oversight), meet people where they are (low-tech options, simple language), support relationships not replace them, and fit AI into existing workflows with minimal disruption.
Bottom line: AI for LEP patients must be sociotechnical—attentive to people, culture, and institutions—not just linguistic.
Paper: http://arxiv.org/abs/2511.07277v1
Paper: http://arxiv.org/abs/2511.07277v1
Register: https://www.AiFeta.com
#AI #Healthcare #HealthEquity #LanguageAccess #LEP #HumanCenteredDesign #DigitalInclusion #Privacy
