UN launches AI-ready data platform

The United Nations announced a partnership with Google to create the UN System Data Commons, a new platform that lets AI systems search and access global statistics using natural language queries.

Replacing outdated data portal

The new system replaces the existing UNData portal, which required users to browse through a traditional database interface. The Data Commons platform supports the Model Context Protocol, allowing AI systems to connect directly to data sources.

AI struggles with data accuracy

A UNICEF benchmark tested six large language models on over 133,000 questions about global development indicators and found an average accuracy score of just 21.2%.

Models tested in benchmark

The study evaluated 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 thousands of development questions.

Models hedge and contradict themselves

About three in five responses did not provide a usable number at all because models hedged their answers. When tested again two days later, models that provided numbers only returned the same answer about half the time.

AI traffic to UNICEF surging

UNICEF saw ChatGPT referrals to its data website rise 67 percent year-over-year between January and September. AI assistants now account for about one in ten visits to the site, which receives over 6 million visits monthly.

26 UN agencies committed

Twenty-six UN entities have committed to the Data Commons platform, with data from nearly 20 available at launch. The UN aims to bring 80 percent of its statistical datasets onto the platform by 2027.

Google funds infrastructure

Google.org provided 2 million dollars in capacity-building funding and technical support. The system is hosted on a UN-governed instance and will eventually be maintained independently by the UN.

AI can trace data to sources

The platform tracks where each statistic originates, allowing people to trace data retrieved by AI systems back to the original UN source. This transparency is crucial as more people rely on AI tools for information.

AI still needs human review

Google demonstrated how AI systems can pull multiple indicators to generate dashboards and analysis automatically. However, humans should always review AI outputs before citing or publishing them, as models can misinterpret nuance.

“Because models can misinterpret nuance, a human should always review the outputs before citing or publishing them.”