Health Equity & Global Access

AI has enormous potential to improve healthcare - but that potential risks being realised primarily in wealthy countries and for affluent populations, widening rather than narrowing global health inequalities. AI diagnostic tools trained primarily on data from European and North American populations often perform poorly for other groups. A dermatology AI trained mostly on lighter skin tones will miss conditions on darker skin. Drug discovery AI focuses on diseases that affect profitable markets rather than neglected tropical diseases. Medical AI tools require infrastructure - reliable electricity, internet connectivity, modern devices, and trained operators - that many healthcare settings in lower-income countries lack. At the same time, AI offers genuine opportunities to extend healthcare access. AI diagnostic tools running on mobile phones could bring specialist-level screening to areas with few doctors. Natural language processing could make medical knowledge accessible in local languages. AI-optimised supply chains could improve the distribution of medicines and vaccines. Realising these benefits requires deliberate effort: training models on diverse, representative data; designing for low-resource settings; making tools affordable and accessible; and involving local healthcare workers in development and deployment. For organisations in the health AI space, building for equity from the start is both an ethical imperative and a path to reaching the largest possible patient population.