Five promising AI-for-good projects
In the current AI and machine learning research explosion are projects that aim to solve some of society’s biggest challenges.
Here are five that could transform lives.
A new approach to help the poor get credit
TrustingSocial is a finance technology project putting AI to work solving the catch-22 the poor face – to get credit, you need to have a credit score and to get a credit score, you have to have had credit. For many, that’s simply impossible. TrustingSocial uses different sources of information to form credit scores, combined with smart algorithms for fraud detection. For example, if somebody reliably pays their phone bill, then they’re probably a reasonable credit risk for moderate sums of money. The program is being driven by Professor Dinh Phung of Monash University, Australia, who is TrustingSocial’s Director of AI Research.
Say goodbye to malaria with a tailored game plan
Eradicating malaria in two years: sounds like a pipe dream, but that’s the aim of Israeli start-up Zzapp. The company’s planning tool analyses environmental and epidemiological factors to tailor insecticide spraying strategies to individual villages and towns. Mobile apps help keep track of house-spraying interventions and guide workers in the field to treat water sources with insecticide at the right time. All data are collated and tracked to highlight undertreated areas or underperforming workers. The project has already run successful pilots in Ghana and Zanzibar, with the latter incorporating drones.
Fresh water for all, even under climate change
Some 1.1 billion people globally don’t have access to fresh water, and by 2025, two-thirds of the world’s population may face water shortages, according to the World Wildlife Fund. Life-sustaining water systems such as rivers and aquifers are becoming too polluted to use or drying up altogether, and climate change is shifting weather patterns, causing droughts and floods. US start-up T2H2O is creating predictive tools that help better manage water supplies, on village, region and country levels, in the face of climate change. T2H2O team leader Jay Bettinger said the technology will also uncover complexities in water management that are yet to be discovered.
Virtual scales detect low birth weight babies
Low birth weight babies are at higher risk of dying in their first year and, for those who survive, lifelong complications such as diabetes. But nearly half of the world’s newborns are not weighed – more so in rural or developing areas lacking accurate weighing devices – and infants that need immediate care often go without. In India, virtual scales developed by the Wadhwani Institute for Artificial Intelligence lets health workers making home health assessment visits use their smartphone to take a short video of a baby. The software then builds 3D models of the infant to estimate its weight, length and head circumference. If a child is deemed severely underweight, the parents are advised to go to the hospital immediately.
Algorithms could ease the refugee crisis
The world is seeing the highest levels of population displacement on record, according to the UN Refugees Agency: of the more than 68 million people that have been forced from home, some 25.4 million are refugees, and these numbers are forecast to rise. Destination countries can be underprepared and overwhelmed by refugee migration, and those seeking refuge can spend months or years in camps before resettlement. To help governments manage refugee movements, the University of California Berkeley and AI For Good Foundation are developing an algorithm to predict migration numbers and patterns, and recommend locations best suited to resettlement.