
What happened
The UN announced Thursday it is working with Google to build the UN System Data Commons on Google's open source Data Commons, replacing UNData and supporting MCP so AI systems can connect directly to UN statistics.
Why it matters
The new platform is designed to make the UN's data AI-ready, with 26 UN entities committed and nearly 20 available at launch, as users increasingly turn to AI tools for answers.
What to watch
Whether the UN's goal of bringing 80% of its statistical datasets onto the platform by 2027 is met hinges on adoption across agencies; watch for the platform's independent maintenance by the UN.
WHO IT HITSUN agency statisticians and data teams who currently maintain UNData portals, as well as developers building AI agents that query global development data, may need to adapt to the MCP-based interface and trace sources back to UN origins.
Summaries like this, every morning.
The UN's move comes after a UNICEF benchmark exposed how poorly leading AI models handle authoritative development statistics. That test, a working paper not yet peer-reviewed, found an average accuracy of 21.2% across six models and showed that even when models gave a number twice, they matched only about half the time. UNICEF also saw a 67% year-over-year rise in visits from ChatGPT answer links between January 1 and September 14, with AI assistants now accounting for about one in 10 visits. The new platform is built on Google's Data Commons, which launched in 2018 and added MCP support last year. Google.org provided $2 million in funding, and the system is hosted on a UN-governed instance intended to be maintained independently by the UN. It aims to bring 80% of the UN system's statistical datasets onto the platform by 2027, with 26 entities committed and nearly 20 available at launch. A Google demonstration showed an AI system using MCP to pull multiple indicators and generate an infographic on the impact of the U.S. President's Emergency Plan for AIDS Relief in Africa. However, as Prem Ramaswami of Google noted, giving AI authoritative data does not make its conclusions authoritative, and a human should always review outputs before citing them. The outcome hinges on whether UN agencies scale adoption and whether the platform's traceability features prove sufficient for users who need citable numbers.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Visa expanded its Agentic Ready program to merchants, payment service providers, acquirers and commerce platfo…

Microsoft is aiming to challenge OpenAI and Anthropic with cheaper, specialized AI models, according to Axios

Visa has outlined a plan for "agentic commerce," a setup where AI agents — software that acts on a user's beha…

Visa's CMO discussed how AI agents are transforming shopping in a WSJ interview

A trial of SRE combined with generative AI for Aeon's DX push surfaced both challenges and possibilities, with…

KDDI's brand communication team now uses Gemini to analyze advertising data, cutting aggregation from roughly…
