The United Nations is working with Google to make its global statistical data easier for artificial intelligence systems to access, as the growing use of AI assistants changes how people search for development and public-sector information.
The organisations have launched the UN System Data Commons, an open platform built on Google's open-source Data Commons technology. The system brings statistics from multiple UN entities into a single interconnected resource and allows users to search the information using natural-language queries.
The platform is designed to replace the existing UNData portal and make official statistics more accessible to both people and AI systems. It also supports the Model Context Protocol, or MCP, an open standard that enables AI agents to connect directly with external data sources.
The move follows concerns about the ability of large language models to reliably retrieve statistical information. According to a UNICEF working paper cited alongside the launch, six major AI models were tested across more than 133,000 responses to questions involving global development indicators.
The models recorded an average accuracy score of 21.2%, according to UNICEF Chief Statistician João Pedro Azevedo. Around three in five responses did not return a usable number. When the same questions were repeated on the same model versions about two days later, answers that provided figures on both occasions returned the same number only around half the time. The study has not yet been peer-reviewed.
The UN System Data Commons is intended to address part of this problem by allowing AI systems to retrieve statistics directly from authoritative datasets rather than relying solely on information contained within a model.
Twenty-six UN entities have committed to participating in the initiative, with data from nearly 20 available at launch. The UN aims to include 80% of the organisation's statistical datasets on the platform by 2027.
Google.org has provided $2 million in capacity-building funding and technical support for the project's core infrastructure. The system is hosted on a UN-governed instance, with plans for the organisation to eventually maintain and scale it independently.
The platform also maintains information about the source of individual statistics, allowing users to trace figures retrieved through AI systems back to the original UN dataset.
AI agents connected through MCP can use the data to combine indicators and generate charts, dashboards, infographics and draft analysis. However, Google has cautioned that access to verified data does not automatically make an AI-generated interpretation authoritative, and users should review underlying sources before publishing or citing critical information.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.