A recent article in The Guardian highlights a remarkable use of AI in Africa: AI-powered pregnancy scans.
The system analyses ultrasound images to detect complications, providing timely and accurate diagnostics previously inaccessible to many expectant mothers in low-resource settings. This technology improves outcomes by enabling early detection of issues and saves lives by preventing stillbirths and other complications.
This story brings to mind another groundbreaking application: Ubenwa. Founded in Nigeria by Charles Onu and his team in 2017, Ubenwa developed an AI-powered mobile app that records the cries of newborns on mobile phones to detect signs of birth asphyxia. Birth asphyxia can lead to severe complications or death if not promptly addressed. Ubenwa's technology uses machine learning to interpret the acoustic features of a baby's cry, providing a non-invasive, cost-effective, and accessible solution that is particularly valuable in regions lacking traditional medical infrastructure.
These innovations exemplify what Clay Christensen described as "disruptive innovation." Often, such breakthroughs arise in resource-constrained environments where traditional solutions are unavailable. Initially, these solutions may seem suboptimal compared to those in resource-rich settings, but they exceed the utility threshold for those with limited resources. Over time, these solutions improve and eventually can outperform the incumbent solutions in resource-rich environments.
There is an understandable emotional desire to hold AI solutions to a higher standard than human performance. While AI must match or exceed human capabilities in resource-rich environments, in resource-constrained settings, AI can provide significant improvements over the status quo. This principle applies not only to mid- and low-income regions but also to richer regions where resources may be limited. AI does not need to be perfect; it needs to be sufficiently effective to make a meaningful difference.
Comments