AI is moving fast—much faster than most governments, civil society groups, or development institutions can realistically keep up with.
Over the past year, we have spoken to partners across our network about the challenges, needs, and opportunities brought by AI, and one message has come through loud and clear: AI’s potential is enormous, but without the right data, governance, and capacity in place, many will be locked out of the benefits.
Distilling the Global Partnership’s role in working with AI in development, today we’re publishing a window into how we’re approaching this rapidly evolving space and inviting partners and funders to shape it with us.
Why now?
AI is no longer theoretical, with governments using chatbots for public services, health ministries using AI to support community workers, and national statistical offices exploring large language models to accelerate data processing.
At the same time, this technology is becoming embedded in people’s lives and societies. Investment is exploding as AI’s potential and how it works in action is understood better, yet profit and visibility—not equity or social good—lead the way in how AI developed and applied. Multilateral bodies, from the G20 to the UN, are working to establish guidelines.
Amid all this momentum, the issue of data is at the core: AI is only as strong as the data it learns from. In many places, that data is missing, incomplete, inaccessible, or not being used with a focus on protecting people’s rights.
What we heard from our network
While headlines fixate on existential threats of AI, the reality in most of the world is much more mundane. Much of the world still lacks access to the internet: up to 30 percent in low-income countries. In the global race to embrace AI technologies, the risks of being left out or left behind are just as salient as the risks of harm. While funders and governments seek to invest in AI for development, addressing the gaps in capacity, skills, governance, and data are paramount to ensuring that new technologies are deployed in ways that are fair and efficient and that benefit people equitably.
Through listening tours, conversations, and surveys across our network of over 700 partners, we heard about consistent concerns and needs:
- People want practical guidance, relevant to their work and contexts, and to learn from others.
- Governments need capacity, as well as tools.
- We must focus on ensuring that AI addresses real-world challenges without widening inequalities.
- Funders want clarity on what meaningful, responsible AI investment can look like.
Our strategy reflects this collective insight.
How we work on AI
AI and emerging technologies is one of our three development challenges. Our contribution focuses on our strengths in data, governance, partnerships, and convening power.
Our approach centers on three roles. The Global Partnership acts as a:
- Learning convenor strengthening institutional capacity to use and govern AI responsibly;
- Solutions broker connecting development challenges with the right partners, data, and tools;
- Coalition builder elevating diverse voices and aligning actors around responsible, inclusive AI.
These roles power an overarching goal of bridging political momentum, technical expertise, and local realities.
What does this mean in practice?
Across our network, powerful examples are already emerging, including:
- A citizen-facing water utility chatbot in Argentina
- AI-supported community health tools in Kenya
- LLM-assisted survey processing in Ghana
- Government partners exploring AI tools for flood preparedness in Sierra Leone
Each of these examples illustrate what could be possible when AI is grounded in real needs and supported by the right data foundations.
What we are inviting you into
This strategy is a starting point for a shared direction of travel. It is a place to open a conversation, and we invite partners, funders, governments, and civil society to partner with us.
By joining forces, together we can shape an AI future that supports progress accelerating and coming along for the ride.
