Duke-NUS Medical School researchers have discussed how artificial intelligence (AI) models can aid clinical decision-making and patient outcome prediction in settings with limited resources, while emphasising the need for governance frameworks and collaborative oversight.

• Adapted AI performance: A transferred AI model accurately distinguished neurological recovery chances after cardiac arrest in a Vietnamese cohort, outperforming the unadapted version.
• Broader application: The press release notes expanded AI use could support diagnostics and decision-making in low- and middle-income countries, citing examples such as pregnancy information chatbots and malaria detection tools.
• Governance emphasis: Researchers propose an international consortium to establish safety, oversight and ethical standards for AI deployment in healthcare.

This summary is based on a media release by Duke-NUS Medical School.