Please see this new report on AI use cases in global health:
I want to support the thrust into AI but can’t shake the parallels to “Blockchain” and “Dark web searches” and “semanitc web” and many other tech trends that are great for fundraising but either under-delivered or were hammers looking for nails.
Thinking about my time in healthcare clinics in Nigeria last year, I have two reflections. First, how in the world can AI touch this facility when electricity is nonexistent, there is no reliable MOH Internet, and no computers? And, then I imagine the look on the head nurse’s face when I start talking AI to her. Couldn’t be less relevant to her world. I feel like this is a real mismatch to what’s actually needed.
For AI or machine-learning algorithms to work, they need:
- High volumes of data - NOPE
- Consistent workflows for data - NOPE
- High quality or at least stable quality data - NOPE
Chasing AI will bring in investment, but it’s being pushed into an environment that is unlikely to be able to use it, more unlikely to be able to sustain it. And every dollar we drag towards AI is dragged away from improving primary care facilities, health infrastructure, etc. My primary example is Tuberculosis in the US versus Mozambique. Both use the same equipment to diagnose; both use the same drugs to treat. Why, then, is TB an epidemic in one country but nearly nonexistent in the other? It certainly isn’t AI. We know what’s needed and I fear the AI train is just the latest in shiny tech pennies we’re grabbing from the wishing well.