UN turns to Google to make its global data ready for AI agents

The United Nations has partnered with Google to create the UN System Data Commons, a new platform designed to make the organization's extensive global statistics more readily accessible to AI systems. This platform, built on Google's open-source Data Commons platform, enables users to search for statistics from across various UN agencies using natural language queries. It replaces the older UNData portal, which required users to browse through a more traditional database interface.

The new platform also supports the Model Context Protocol (MCP), which allows AI systems to connect directly to external data sources. With users increasingly relying on AI tools for answers, the UN found that many systems struggle to reliably surface authoritative data. A UNICEF benchmark tested accuracy of six large language models answering questions about global development indicators, with an average accuracy score of just 21.2%.

About two-thirds of the model responses did not provide a usable number, often due to models hedging their answers. However, when the same questions were re-run on the same model versions a couple of days later, the models provided the same number only about half the time.

This new UN platform addresses these issues, as it keeps track of where each statistic comes from, allowing users to trace data retrieved by an AI system back to the original UN source. UNICEF has also observed a significant increase in traffic from generative AI assistants to its data website, with visits from these assistants rising by 67% year-over-year. The platform is currently hosted on a UN-governed instance and aims to bring 80% of the UN system's statistical datasets onto the platform by 2027.