--- title: UN turns to Google to make its global data ready for AI... description: The United Nations launched the UN System Data Commons, built on Google's open source Data Commons platform, to make its global statistics accessible to AI... canonical: https://daily.dev/posts/un-turns-to-google-to-make-its-global-data-ready-for-ai-agents-mlquabn89 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: UN turns to Google to make its global data ready for AI agents | daily.dev og:description: The United Nations launched the UN System Data Commons, built on Google's open source Data Commons platform, to make its global statistics accessible to AI... og:url: https://daily.dev/posts/un-turns-to-google-to-make-its-global-data-ready-for-ai-agents-mlquabn89 og:image: https://api.daily.dev/og/posts/MLqUABN89.png og:image:alt: UN turns to Google to make its global data ready for AI agents og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # UN turns to Google to make its global data ready for AI agents **[TechCrunch](https://daily.dev/sources/tc)** · 4 min read · 0 upvotes · 0 comments ## Summary The United Nations launched the UN System Data Commons, built on Google's open source Data Commons platform, to make its global statistics accessible to AI agents via natural-language queries and the Model Context Protocol (MCP). It replaces the older UNData portal. The move follows a UNICEF benchmark testing six LLMs (GPT-4o, GPT-4o-mini, Claude Sonnet 4.5, Haiku 4.5, Gemini 2.5 Flash, Gemini 2.0 Flash) across 133,000+ questions, finding only 21.2% average accuracy on global development indicators, with about three in five responses giving no usable number and inconsistent answers when re-run. Google.org provided $2 million in funding; 26 UN entities have committed, with nearly 20 datasets live at launch and a goal of 80% coverage by 2027. The platform tracks data provenance so AI-generated answers can be traced to original sources, though Google stresses human review remains necessary. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How accurate were GPT-4o, Claude Sonnet 4.5, and Gemini 2.5 Flash at retrieving UN global development statistics? A UNICEF benchmark testing six large language models (GPT-4o, GPT-4o-mini, Claude Sonnet 4.5, Haiku 4.5, Gemini 2.5 Flash, and Gemini 2.0 Flash) across more than 133,000 responses found an average accuracy of just 21.2% on global development indicator questions. About three in five responses gave no usable number at all, often due to hedging, and when re-run two days later on the same model versions, matching numeric answers were identical only about half the time. _Anyone evaluating LLM reliability for factual data retrieval can track findings like these on daily.dev._ ### What is the UN System Data Commons and how does it use MCP? The UN System Data Commons is a new platform built on Google's open source Data Commons, replacing the older UNData portal, that lets users query UN statistics using natural language and supports the Model Context Protocol (MCP) so AI agents can connect directly to the data. It tracks provenance for each statistic, letting users trace AI-retrieved figures back to their original UN source. Google.org contributed $2 million toward building the platform, and 26 UN entities have committed to it, aiming for 80% of UN statistical datasets by 2027. _Developers wiring AI agents to authoritative data sources can follow MCP integration news on daily.dev._ ## Similar posts on daily.dev - [The New Data Commons MCP Server Unlocks a Wealth of Public Datasets for AI Developers](https://daily.dev/posts/the-new-data-commons-mcp-server-unlocks-a-wealth-of-public-datasets-for-ai-developers-uujgr3qvn) · InfoQ · 1 upvotes · 0 comments - [Data Commons Hosted MCP: Zero-Install Public Data for AI](https://daily.dev/posts/data-commons-hosted-mcp-zero-install-public-data-for-ai-takp8rijy) · Google Developers · 1 upvotes · 0 comments - [Google is powering a new US military AI platform](https://daily.dev/posts/google-is-powering-a-new-us-military-ai-platform-f7why4vco) · The Verge · 0 upvotes · 0 comments - [EU forces Google to share its toys with the other AI and search kids](https://daily.dev/posts/eu-forces-google-to-share-its-toys-with-the-other-ai-and-search-kids-xhdlkxy9y) · The Register · 1 upvotes · 1 comments --- Tags: [#mcp](https://daily.dev/tags/mcp), [#rag](https://daily.dev/tags/rag) [View this post on daily.dev](https://daily.dev/posts/un-turns-to-google-to-make-its-global-data-ready-for-ai-agents-mlquabn89) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"UN turns to Google to make its global data ready for AI agents","url":"https://daily.dev/posts/un-turns-to-google-to-make-its-global-data-ready-for-ai-agents-mlquabn89","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/un-turns-to-google-to-make-its-global-data-ready-for-ai-agents-mlquabn89"},"datePublished":"2026-09-17T20:02:02.017Z","dateModified":"2026-09-18T13:28:06.598Z","description":"The United Nations launched the UN System Data Commons, built on Google's open source Data Commons platform, to make its global statistics accessible to AI...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/1d70e3e21709d32cfcbcdd3b748c5b9d?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/1d70e3e21709d32cfcbcdd3b748c5b9d?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"TechCrunch","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"TechCrunch","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/tc","url":"https://daily.dev/sources/tc"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/un-turns-to-google-to-make-its-global-data-ready-for-ai-agents-mlquabn89","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"mcp,rag","timeRequired":"PT4M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"TechCrunch","item":"https://daily.dev/sources/tc"},{"@type":"ListItem","position":3,"name":"UN turns to Google to make its global data ready for AI agents"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/un-turns-to-google-to-make-its-global-data-ready-for-ai-agents-mlquabn89#faq","mainEntity":[{"@type":"Question","name":"How accurate were GPT-4o, Claude Sonnet 4.5, and Gemini 2.5 Flash at retrieving UN global development statistics?","acceptedAnswer":{"@type":"Answer","text":"A UNICEF benchmark testing six large language models (GPT-4o, GPT-4o-mini, Claude Sonnet 4.5, Haiku 4.5, Gemini 2.5 Flash, and Gemini 2.0 Flash) across more than 133,000 responses found an average accuracy of just 21.2% on global development indicator questions. About three in five responses gave no usable number at all, often due to hedging, and when re-run two days later on the same model versions, matching numeric answers were identical only about half the time. Anyone evaluating LLM reliability for factual data retrieval can track findings like these on daily.dev."}},{"@type":"Question","name":"What is the UN System Data Commons and how does it use MCP?","acceptedAnswer":{"@type":"Answer","text":"The UN System Data Commons is a new platform built on Google's open source Data Commons, replacing the older UNData portal, that lets users query UN statistics using natural language and supports the Model Context Protocol (MCP) so AI agents can connect directly to the data. It tracks provenance for each statistic, letting users trace AI-retrieved figures back to their original UN source. Google.org contributed $2 million toward building the platform, and 26 UN entities have committed to it, aiming for 80% of UN statistical datasets by 2027. Developers wiring AI agents to authoritative data sources can follow MCP integration news on daily.dev."}}]} ```