A2i & GiveDirectly joined forces to combat poverty by the usage of AI

a2i, in partnership with GiveDirectly, a nonprofit that delivers unconditional cash transfers to impoverished communities, hosted an international learning event in the capital.  

a2i and GiveDirectly have been working closely together with funding from USAID & Google.org and technical assistance from the University of California Berkeley, for the last year to utilise machine learning and phone data to improve the targeting of humanitarian aid in Bangladesh. They have remotely identified vulnerable Rohingya refugee hosts and impacted communities and delivered cash transfers through mobile money. 

This transformational approach predicts the poverty status of individual mobile phone users based on their phone usage patterns. The project aligns with the government's Smart Bangladesh Vision by harnessing technology to quickly and efficiently identify and assist those most in need during emergencies.

The event also marked the launch of an innovative climate crisis-responsive cash project to address key questions relevant to anticipatory action efforts globally in partnership with the University of Oxford and Google Flood Forecasting Initiative, including generating further evidence on providing anticipatory vs. post-disaster cash assistance.

Attendees at the event included key stakeholders from the public and private sectors, development partners, and academic institutions. "By harnessing technology to optimise humanitarian aid delivery, we are advancing the government's vision of leveraging AI to drive positive change,  enhance efficiency across all facets of governance, and contribute to broader national development objectives," said Mollah Mizanur Rahman, Joint Project Director of a2i.

"In the wake of COVID-19, GiveDirectly brought together a team in Togo to use machine learning and mobile phone data to remotely identify, enrol, and pay a total of USD 10 million to 138,000 Togolese nationwide as emergency relief. This approach caught the attention of a2i, who were keen to replicate it in the Bangladesh context and began discussing a collaboration with GiveDirectly here." Said Lamia Rashid, Regional Director of GiveDirectly.  

Key discussions at the event centred on the broader social protection landscape, emphasising the potential of AI-driven solutions to target aid and manage climate risks efficiently. Stakeholders exchanged insights on the importance of collaboration between government, academia, NGOs, and the private sector in driving impactful initiatives.