How AI will change global development

For artificial intelligence to advance rather than undermine global development work, leaders need to shift from supporting experiments to focusing on infrastructure, systems, and implementation. That’s according to a panel of experts who joined Devex for a conversation as part of a Devex Pro Week event that asked: How will AI change development? To ensure the benefits of AI are experienced around the world, the global development community must focus on closing the “digital divide,” or ensuring broader access to connectivity; strengthening data architecture and governance; and building digital public infrastructure, or digital identity, payments, and data exchange that allow people to access social services. “We are talking about digital transformation,” said Prateek Sibal, program specialist at the Digital Policies and Digital Transformation Section of the United Nations Education, Scientific, and Cultural Organization. “We often just limit ourselves to digital and forget transformation and move from one hype to the other. But that transformation is where the real work needs to happen.” Sibal was joined by Priya Vora, the managing director of the Digital Impact Alliance, which is working to extend the benefits of trusted digital technology around the world; Zameer Brey, the head of technology diffusion in the global health division of the Bill & Melinda Gates Foundation; and Lavina Ramkissoon, an adviser to the African Union on technology and AI, for a wide-ranging conversation that also covered how to elevate global voices, build an evidence base for responsible AI, and avoid biases and misuse. Sibal provides policy advice on AI to governments and develops programs to support policymakers and civil society organizations in low- and middle-income countries. Last month, policymakers from sub-Saharan Africa and India who were gathered for a workshop in Kigali, Rwanda, shared how the problems they face with AI have mostly to do with coordination. Governments have a hard time developing AI strategies that will get the buy-in they need across departments — from creating an action plan to dealing with data and interoperability to ensuring there is adequate funding, Sibal said. AI depends on good data, which is often not available in many of the contexts that could benefit most from these emerging tools, Vora said. “I think the key aspect of our work right now that we think directly impacts whether AI is broadly beneficial or not is our work in strengthening data architecture and data governance systems with country partners.” Vora and fellow panelists spoke to additional challenges that stand in the way of AI adoption in low-resource settings, including internet access and access to reliable power. This month, the Gates Foundation will announce which projects will receive grants from a $3 million request for proposals around the use of large language models in low- and middle-income countries. There was so much interest, with 1,300 applications from 103 countries, that the foundation may increase the funding that was originally made available, Brey said. The Gates Foundation is committed to involving low-income countries in the creation of AI tools, not just through these grants, but through an external advisory group focused on responsible AI, as well as a new grant to an Africa-based network of AI ethicists and gender experts that will serve as a resource to foundation staff as well as grantees. “If we want to see AI as a tool for change, as a force multiplier, I think we’ve got to invest in building out that evidence base so that we can outlive the hype that’s with us now,” Brey said.