UN Partners with Google to Enhance AI Access to Global Data

The United Nations has announced a partnership with Google to enhance the accessibility of its global statistics for AI systems. This initiative, known as the UN System Data Commons, is built on Google's open-source Data Commons platform. It aims to facilitate the use of natural-language queries to search for statistics from various UN agencies, replacing the older UNData portal that required traditional database navigation.

The new platform supports the Model Context Protocol (MCP), which allows AI systems to connect directly to external data sources. This development comes in response to challenges identified in a UNICEF benchmark study, which assessed the accuracy of AI responses to questions about global development indicators. The study found that AI systems, including OpenAI's GPT-4o and Google's Gemini models, struggled to provide reliable data, with an average accuracy score of just 21.2% across over 133,000 responses.

João Pedro Azevedo, UNICEF's chief statistician, highlighted that many AI models failed to provide usable numbers, often hedging their answers. Even when models did provide numbers, they returned the same result only about half the time when asked the same question again two days later. This study is currently being prepared for journal submission, with plans to release its methodology, code, and data alongside the paper.

UNICEF has also observed a significant increase in traffic from generative AI assistants to its data website, which is among the agency's most popular, receiving over 6 million visits monthly. Visits from users clicking links in ChatGPT answers rose 67% year-over-year, accounting for 6.4% of all sessions this year. Overall, AI assistants now account for approximately 10% of visits.

The UN reported that 26 of its entities have committed to the Data Commons, with data from nearly 20 available at launch. The goal is to have 80% of the UN system's statistical datasets on the platform by 2027. Shantanu Mukherjee, acting director of the UN Statistics Division, emphasized the scale, scope, and flexibility of this new system, which connects numerous agencies across the UN system and prepares their data for AI use.

Google.org has contributed $2 million in capacity-building funding and technical support to establish the UN System Data Commons. This collaboration represents a significant step forward in making global data more accessible and usable for AI applications, potentially improving the accuracy of AI-generated responses to queries about global development.

The partnership also aims to address the growing reliance on AI tools for data retrieval. As more users turn to AI for information, the need for accurate and reliable data becomes critical. The UN System Data Commons is expected to play a crucial role in ensuring that AI systems can access authoritative data, thereby enhancing the quality of information provided to users worldwide. This initiative not only benefits AI developers but also supports global development efforts by providing more accurate data insights.

Related reading and gear: Media literacy book.

Related reading and gear: Desk phone stand.