UN Partners With Google to Make Global Statistics Accessible to AI Systems
A UNICEF benchmark found six leading AI models averaged just 21.2% accuracy on global development questions, with roughly three in five responses lacking a usable number. The UN’s new platform uses Google’s MCP-ready infrastructure to improve data access.
The United Nations announced in September 2026 that it is working with Google to make its global statistical data easier for AI systems to access and use. The new platform, called the UN System Data Commons, is built on Google’s open-source Data Commons infrastructure and replaces the existing UNData portal, which relied on a traditional database browsing interface.
The new system allows users to search statistics from across UN agencies using natural-language queries and supports the Model Context Protocol (MCP), a standard that enables AI systems to connect directly to external data sources. Google demonstrated how an AI system connected via MCP could retrieve multiple UN indicators and generate dashboards, charts, and written analysis automatically — including an infographic on the impact of U.S. AIDS relief programs in Africa using HIV, mortality, and life expectancy data.
The shift was driven in part by a UNICEF benchmark study that tested six large language models — including OpenAI’s GPT-4o and GPT-4o-mini, Anthropic’s Claude Sonnet 4.5 and Haiku 4.5, and Google’s Gemini 2.5 Flash and Gemini 2.0 Flash — across more than 133,000 responses to questions about global development indicators. The models returned an average accuracy score of just 21.2%, and roughly three in five responses failed to provide a usable number at all. The study has not yet been peer-reviewed.
UNICEF also reported a 67% year-over-year rise in visits to its data website from ChatGPT referrals, with AI assistants now estimated to account for about one in ten of the site’s more than six million monthly visits.
Google.org contributed $2 million in funding and technical support to establish the platform’s core infrastructure. Twenty-six UN entities have committed to the Data Commons, with data from nearly 20 available at launch. The UN aims to bring 80% of its statistical datasets onto the platform by 2027.
The platform tracks the source of each statistic so data retrieved by an AI system can be traced back to the original UN source. However, Prem Ramaswami, who leads Google’s Data Commons team, cautioned that authoritative data does not guarantee authoritative conclusions: “Because models can misinterpret nuance, a human should always review the outputs before citing or publishing them.”