The United Nations announced on Thursday that it is partnering with Google to create a new platform that makes the UN’s global statistics easier for AI systems and people to access. The system, named UN System Data Commons, is built on Google’s open-source Data Commons platform and replaces the older UNData portal.
What the new system does
UN System Data Commons lets users search UN agency statistics with natural-language queries rather than the traditional database interface of UNData. The platform also supports the Model Context Protocol (MCP), a standard designed to allow AI systems to connect directly to external data sources and retrieve statistics and provenance information.
Why this matters now
AI tools are increasingly used to answer questions about global indicators, but many large language models struggle to reliably surface authoritative numbers. João Pedro Azevedo, UNICEF’s chief statistician, told reporters that a UNICEF benchmark testing six large language models across more than 133,000 responses to questions about global development indicators produced an average accuracy score of just 21.2%.
The models in that test included 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. Azevedo said roughly three in five responses did not provide a usable number, often because the models hedged. Moreover, when the same questions were rerun about two days later, model runs that returned a number both times provided the identical number only about half the time.
UNICEF described the work as a working paper being prepared for journal submission that has not yet been peer-reviewed; the organization plans to release methodology, code, and data alongside the paper.
Rising traffic from AI assistants
UNICEF has also seen a sharp increase this year in visits referred by generative AI assistants. Its data website receives more than six million visits per month and is among the agency’s most popular sites. Between January 1 and September 14, referrals from links in ChatGPT answers rose 67% year over year, Azevedo told TechCrunch. Those referrals made up 6.4% of sessions this year, and UNICEF estimates AI assistants now account for about one in ten visits overall.
Scope, funding and targets
The UN says 26 of its entities have committed to the Data Commons, with data from nearly 20 available at launch. The organization aims to bring 80% of the UN system’s statistical datasets onto the platform by 2027. Google.org provided $2 million in capacity-building funding and technical support to establish the platform’s core infrastructure.
Prem Ramaswamy, who leads Google’s Data Commons team, said the system runs on a UN-governed instance and is intended to be maintained, operated, and scaled independently by the UN over time. He added that Google used a “train-the-trainer” approach during rollout and has already seen UN teams ramp up quickly.
Shantanu Mukherjee, acting director of the UN Statistics Division, said the platform represents a major advance in scale, scope and flexibility, connecting many agencies across the UN system for the first time and making the data "AI-ready."
Provenance and demonstrations
The UN platform tracks where each statistic comes from so data returned by AI can be traced back to the original UN source. Azevedo emphasized the importance of that traceability as more people rely on AI tools to find and interpret information.
Google demonstrated how an AI connected to UN data via MCP could assemble multiple indicators to produce dashboards, charts, and written analysis without a user manually combining datasets. In one demo, Google asked an AI to find the impact of the U.S. President’s Emergency Plan for AIDS Relief (PEPFAR) in Africa. The system identified relevant UN statistics on HIV infections, AIDS mortality and life expectancy, and used them to generate an infographic.
Human oversight remains necessary
Providing AI systems with authoritative data sources does not automatically make their outputs authoritative. As Prem Ramaswamy noted, models can misinterpret nuance, so a human should always review outputs before citing or publishing them.
Overall, the UN System Data Commons aims to make UN statistics more directly usable by modern AI ecosystems while preserving provenance and creating a platform the UN can ultimately govern and sustain.
