UN partners with Google to make global data accessible to AI agents

The United Nations has announced a partnership with Google to launch the "UN System Data Commons," a new system designed to make its vast collection of global statistics more accessible to AI systems. Built on Google’s open-source Data Commons platform, the new system allows users to search for statistics from across UN agencies using natural-language queries, replacing the traditional database interface of the existing UNData portal. The platform also supports the Model Context Protocol (MCP), enabling AI systems to connect directly to external data sources. This initiative addresses the challenge of AI systems struggling to reliably surface authoritative data. João Pedro Azevedo, UNICEF’s chief statistician, noted that a benchmark of six large language models—including OpenAI’s GPT-4o, Anthropic’s Claude Sonnet 4.5, and Google’s Gemini 2.5 Flash—produced an average accuracy score of just 21.2% when answering questions about global development indicators. Azevedo reported that about three in five responses failed to provide a usable number, often due to the models hedging their answers. Furthermore, when the same questions were repeated two days later, models provided identical numbers only about half the time. UNICEF plans to release its methodology, code, and data alongside a working paper currently being prepared for journal submission. The move comes as the agency sees a sharp rise in traffic from generative AI assistants to its data website, which currently receives over 6 million visits per month.