UN and Google build an AI-ready system for global statistics
The United Nations is working with Google on the UN System Data Commons, a platform designed to make statistics from UN agencies easier for people and AI systems to find. Built on Google’s open-source Data Commons, it accepts natural-language queries and replaces the older UNData portal’s mainly database-style interface. It also supports the Model Context Protocol, allowing AI systems to connect directly to external data sources.
The project follows a UNICEF test of six large language models that examined more than 133,000 answers about global development indicators. The models achieved an average accuracy rate of 21.2%, while about three in five responses failed to provide a usable number. Among answers that included a number in two repeated tests, the result matched only about half the time. UNICEF described the research as a working paper that has not been peer-reviewed.
Twenty-six UN entities have committed to the platform, with data from nearly 20 available at launch. The UN aims to add 80% of its statistical datasets by 2027. Google.org supplied $2 million in funding and technical support. The system records each statistic’s original source, but Google’s Prem Ramaswamy said people should review AI-generated interpretations before publishing them.
Written by Sakhanda Wire AI from reporting by TechCrunch. Facts, quotes and full context are in the original article below.