Two people holding mobile phones in Kigali, Rwanda to illustrate the digital transition to a cashless economy using mobile payments. Photo by Godong/Universal Images Group via Getty Images.
Governing with AI in low and middle-income countries (LMICs)
For low- and middle-income country governments, AI provides a major opportunity: to strengthen public services and make better use of fragmented public data. The question is whether governments can shape its use in ways that are responsible, effective, and grounded in public purpose.
Please note that you must request an invitation to attend this event.
Governing with AI in LMICs Event Graphic
Artificial Intelligence (AI) could help governments improve health system planning, strengthen climate adaptation, support food-security forecasting, and reduce administrative bottlenecks. But these gains are not automatic. Poorly designed systems can reproduce bias, deepen exclusion, weaken accountability, and erode public trust. Models built primarily on data from high-income countries may perform poorly in other settings. Weak data systems, limited institutional capacity and inadequate safeguards can mean that errors go undetected until they affect the people least able to challenge them.
This event will bring together a select group of ministers, senior public officials, technology leaders, researchers, funders and policy experts for a high-level discussion on how AI can be used responsibly, at scale, and in ways that improve public outcomes.
Convened by the International Growth Centre (IGC), the World Bank and the LSE Data Science Institute, the conference will focus on the institutional, technical, and political choices that determine whether AI strengthens public institutions or bypasses them. Discussions aim to move beyond broad claims about AI’s potential and examine what governments need to get right: data governance, public-sector capability, accountability, human rights protections, local adaptation, and the sequencing of reforms.
The event will explore questions including:
- What public-sector problems is AI genuinely well suited to solve?
- What institutional capabilities do governments need to deploy AI responsibly and at scale?
- How can AI support better decisions in public health and climate adaptation?
- How can governments manage risks of bias, exclusion, and error when public data systems are fragmented or incomplete?
- What does a human-rights approach to AI governance mean in practice for developing-country governments?
- How can AI build human capability and strengthen public institutions, rather than erode local capacity?
- What role should international organisations, technology companies, researchers and funders play in supporting responsible AI adoption?
Through high-level panels, practical case studies and focused discussion, the event will examine how governments can build the capabilities, partnerships and guardrails needed to deploy AI in the public interest. Its aim is to chart a more practical path for AI in government: one that is ambitious about technology, realistic about institutional constraints, and clear about the rights and public outcomes it is meant to serve.
This event is invite-only. You can request an invite here, and if you would like to discuss broader collaboration, please reach out to Shahrukh Wani at [email protected].
We are delighted to share our preliminary agenda below and an initial selection of confirmed speakers. Please note that additional speakers and the full programme lineup will be announced in the coming weeks.
Keynote: Why AI matters for the state
How the UK government built an in-house AI unit
Session 1: Building AI foundations in government systems
Paula Ingabire
Minister of Information Communication Technology and Innovation of Rwanda, Rwanda
Salima Monorma Bah
Minister of Communications, Technology and Innovation, Sierra Leone
Kusobile Kamwambi
Head of the Presidential Delivery Unit (PDU), Government of Zambia
Shahrukh Wani
Head of Data and AI Policy Partnerships, International Growth CentreBefore asking what AI can do, governments need to ask what they are actually trying to fix. This session focuses on problem identification and state capacity: where AI adds genuine value in core government functions, what institutional conditions make deployment work, and what governments need to build before they can use AI responsibly at scale.
Coffee break
Hall of innovations - AI tools on showcase.
How the World Bank is forecasting food insecurity with AI
Session 2: Climate adaptation
Jonathan Pershing
Dean, Professor in Practice, LSE’s Global School of Sustainability
Joanne Ke Edelman
Associate Director, Global Development Incubator
Philip Frost
Chief Technology Officer, TomorrowNow.org
Rose M. Mutiso
Founder and Executive Director, African Tech Futures Lab
Tim Dobermann
Research Director, International Growth CentreOf all the domains where AI might support better government decisions, climate adaptation has some of the clearest use cases and some of the hardest constraints. The decisions are high-stakes and often irreversible. The data is frequently fragmented, localised, or absent. This session examines how AI can support better anticipation, prioritisation, and response - and what it takes to make those capabilities reliable in the settings where they are needed most.
Lunch
Curates group discussions centred on key themes
How the Government of Ukraine is deploying AI
In collaboration with Deloitte and the FCDO
Session 3: Public Health
Aleta Starosta
Head of the Zambia Evidence Lab (ZEL), International Growth Centre
Barbara-Chiara Ubaldi
Senior Director, Client Engagement and Delivery, Tony Blair Institute for Global Change
Charlotte Watts
Executive Director, Solutions, Wellcome
Peter Funsani
Acting Director, Strategic Planning and Information Management, Zambia National Public Health Institute (ZNPHI)Health is the domain where AI’s potential to improve lives is most direct, and where failures of governance, bias, or data protection cause the most harm. This session examines how AI can support better health-system planning, resource targeting, and performance monitoring in settings where resources are scarce and need is acute. It also examines the harder questions: how governments protect sensitive health data, maintain public trust, uphold rights protections, and ensure that AI supports health workers and builds institutional capacity rather than bypassing it.
Coffee break
Hall of innovations - AI tools on showcase
AI-enabled labour-market analytics in Ethiopia
In collaboration with Tabiya
Session 4: Data, bias, and local adaptation
The final session steps back to examine the cross-cutting risks and design choices that shape AI outcomes across sectors. The conversation will focus on how AI systems are trained, deployed, and governed - and how weak institutional safeguards can undermine accountability and human rights even when intentions are good.