Governments worldwide are adopting AI in public services faster than they can build the data infrastructure those services depend on. A vaccination campaign, steered by incomplete records, will reach the wrong districts. A programme meant for households in poverty cannot find the homes missing from the data. At the same time, in many countries, people are already using general-purpose AI assistants to find public information and navigate services, whether or not their governments have adopted the technology themselves.
Data Foundations for AI, a joint effort launched today in New York, is built to change this. Delivered by the Global Partnership for Sustainable Development Data (the Global Partnership) and the Open Data Institute (ODI), it helps to build the institutional, technical and governance conditions that let data be managed, shared and used well, so that AI in public services is built on data infrastructure that can be trusted.
Data Foundations for AI will work to strengthen data governance and policy, build public data infrastructure for AI, and contribute to lasting skills and capabilities inside public institutions. Countries start from different places, from digitising basic records to testing AI in live services, and collaboration will work across that spectrum, tailoring services to meet needs. Governments will maintain control of their nationally important data throughout, and countries and institutions just beginning to work with AI will especially benefit from being able to build foundations well from the start rather than having to retrofit better solutions around systems already in place.
Data Foundations for AI's first in-country delivery is under way:
- In Chile, where it has helped to close the implementation gap facing its new data protection law and separate draft AI bill. The work has involved identifying the technical infrastructure, data standards, and institutional coordination needed to operationalise these laws, and recommending a health sector regulatory sandbox to test them.
- In Sierra Leone, where it is working with the Ministry of Agriculture and Food Security to strengthen agricultural data governance and skills in support of Feed Salone, the national strategy for food security, so that AI-supported decisions on farming and food can draw on data the government trusts and owns.
We founded the ODI to open up public data. The results have had widespread social and economic benefits, and with AI the opportunities from doing so only increase. Data is the bedrock of AI, and realising these opportunities depends on getting the foundations right. Public services can use the power of data for the good of their own citizens, and if they do, it will build a future where the benefits of AI are felt by everyone.
Co-founder of the Open Data Institute and inventor of the World Wide Web
The UN Foundation is proud to host Data Foundations for AI, a collaboration among leading experts in the field that has the potential to be game-changing for improving the quality and governance of the data behind public service. There is no path to a peaceful and prosperous world without making AI work for the public good.
President and CEO, United Nations Foundation
There is much talk of sovereign AI, and if countries are serious about this, that capability will be built off the back of the data they hold and govern. That means knowing where data comes from - its provenance, when, how and why it was generated, the terms under which it can be used, and building the skills to do this inside a country's own institutions. A government that controls its data in this way is not tied to whichever supplier it started with and can switch models or providers as better ones emerge while holding on to the data assets it has built. The governments laying these foundations now are the ones that will have freedom of action in the future.
Co-founder and Executive Chair of the Open Data Institute
For governments to use AI, they have to be able to trust the data it is built on. Our plan is to strengthen national coordination, integrate other data and embed agricultural data in the national work plan and budget, so that we are funding this sustainably rather than depending only on external support.
Deputy Director, Planning Evaluation Monitoring and Statistics Division, Ministry of Agriculture and Food Security, Government of Sierra Leone
Announcing the initiative, Jenna Slotin, CEO, The Global Partnership said:
Governments across low- and middle-income countries are moving ahead with AI on their own terms. They want data systems they control, built around their priorities, from food security to health. Through our network of more than 700 partners, Data Foundations for AI starts from where each country is, and works towards public services that reach people directly.
For thirteen years the ODI has built the standards, tools and skills that make data ready for use, and the lesson of that work is that technical solutions succeed only when institutions adopt and sustain them. This partnership closes a gap neither of our organisations could close alone, combining the ODI’s technical depth with the Global Partnership's reach within governments worldwide. Together we can build data foundations that are nationally owned and still working long after any single project ends.
CEO, Open Data Institute
Data Foundations for AI is being launched in New York at the annual board meeting of The Global Partnership and during the 81st session of the UN General Assembly (UNGA 81).
