The United Nations announced Thursday that it is working with Google to make its vast collection of global statistics easier for AI systems to access and use, replacing the aging UNData portal with a new platform that speaks the language of AI agents.
Why This Matters: AI Gets Global Data Wrong 79% of the Time
The driving force behind the partnership is a sobering UNICEF benchmark. The agency tested six leading large language models across more than 133,000 responses to questions about global development indicators and found an average accuracy score of just 21.2%.
The test covered OpenAI GPT-4o and GPT-4o-mini, Anthropic Claude Sonnet 4.5 and Haiku 4.5, and Google Gemini 2.5 Flash and Gemini 2.0 Flash. About three in five responses did not provide a usable number at all, often because the models hedged their answers. Worse, when the same questions were retested about two days later, models that provided a number both times returned the identical number only about half the time.
The findings underscore a growing problem: as people increasingly turn to AI tools for information, the authoritative data those tools surface is frequently wrong, missing, or inconsistent.
The UN System Data Commons
The new platform, called the UN System Data Commons, is built on Google open-source Data Commons platform and replaces the existing UNData portal, where users had to browse through a traditional database interface. The new system supports the Model Context Protocol, a standard that allows AI systems to connect directly to external data sources.
Twenty-six UN entities have committed to the platform, with data from nearly 20 available at launch. The UN aims to bring 80% of the UN system statistical datasets onto the platform by 2027.
The platform keeps track of where each statistic comes from, so people can trace data retrieved by an AI system back to the original UN source. Google demonstrated how an AI system connected through MCP could pull together multiple indicators and generate dashboards, charts, and written analysis without a user manually finding and combining datasets.
Traffic Shift and Financial Backing
UNICEF has seen a sharp rise in traffic from generative AI assistants to its data website, which receives more than 6 million visits a month. Visits from users clicking links in ChatGPT answers rose 67% year-over-year between January and September, with AI assistants now accounting for about 1 in 10 visits.
Google.org provided $2 million in capacity-building funding and technical support. The system is hosted on a UN-governed instance and is intended to eventually be maintained and scaled independently by the UN.
Caveat: Authoritative Data Does Not Equal Authoritative Conclusions
Google Prem Ramaswami cautioned that giving an AI system authoritative data does not necessarily make its conclusions authoritative. Because models can misinterpret nuance, a human should always review the outputs before citing or publishing them.
