01
MultilingualCoverage
Evaluation covering languages not yet widely included, especially low-resource languages.
Building Global Evidencefor Generative AIin Health Care
- Medical
- AI
- Global
- Initiative
- Collaboration
MAGIC Is An International Network Dedicated To Building A Global Evaluation Ecosystem For Generative AI In Health Care
By bringing together health care professionals, AI researchers, industry stakeholders, policymakers, and local domain experts, MAGIC aims to generate trustworthy and globally representative evaluation evidence, providing the global health community with a shared direction and practical pathway for the responsible deployment of LLMs and other GenAI technologies.
Global Health LLM Evaluation Benchmark
A shared framework for globally representativeevidence across three essential dimensions.
02
Regional DiseaseBurden
Evaluation grounded in region-specific disease spectrums and comorbidity patterns.
03
SocioculturalAdaptability
Evaluation of whether AI can produce reliable responses within local sociocultural contexts.
This is a continuously updated benchmark that adopts a modular design to continuously incorporate additional evaluation resources.
Reporting Standard for Global LLM Deployment
Clarify the scope of application and limitations of GenAI to guideits responsible deployment across diverse health care settings.
Illustrative, Not Exhaustive
Intended Use
Target Users
Health Care Settings
Evaluation Data
Performance and Safety
Scope and Limitations
As technology, evaluation methods, clinical contexts, and deployment needs continue to evolve, additional dimensions will be incorporated.
Join MAGIC
MAGIC Needs Global Participation
Contribute Local Perspectives
Share local practice experience in language, disease patterns, and sociocultural context.
Co-Build Evaluation Resources
Collaborate in constructing evaluation resources and continuously refine them.
Shape Global Standards
Advance responsible model deployment and reporting standards globally.
Questions or collaboration ideas?liu.nan@duke-nus.edu.sg