13 March, 2026

Artificial intelligence is rapidly becoming part of how health systems operate. But much about it remains unknown – and the places where it could make the greatest difference often have the least evidence about how well it works.  

 

The new Evidence for AI in Health (EVAH) initiative, led by Wellcome alongside the Gates Foundation and the Novo Nordisk Foundation, aims to close this gap by generating real‑world evidence in the settings that need it most.  

  

Despite deep expertise and ongoing innovation across low- and middle-income countries (LMIC), there is a persistent shortfall in evaluation. A  recent Lancet study  highlighted this gap: Of the 86 identified randomized clinical trials conducted on AI health tools around the world between 2018 and 2023, only four took place in LMICs.  

 

EVAH plans to address this by strengthening local capacity to assess new technologies and by investing directly in evidence generation. Local validation is essential because it helps identify risks early and prevents AI tools from creating or worsening health inequalities. 

  

The initiative will focus on a range of tools that can be used in primary and community healthcare settings – from models that can predict disease to vision tools that can scan X-rays and scans to large language models that support decision making.   

 

We see this as the type of investment that might not always make headlines, but could play a vital role in driving better health outcomes. It addresses a critical knowledge gap – how AI tools actually perform in real health settings in LMICs. And this means it could play a huge role in unleashing the enormous potential of AI to make a difference in places often facing critical workforce and infrastructure challenges.  

By Tulika Agarwal