On Thursday, the United Nations unveiled the launch of the UN System Data Commons, a collaborative initiative with Google. This system, built on Google’s open-source Data Commons platform and compatible with the Model Context Protocol (MCP), enables users to search for data from UN agencies using natural language. Simultaneously, AI can directly access and leverage authoritative data sources. Tests conducted by UNICEF indicated that the average accuracy rate of six leading large language models in answering questions about global development indicators was a mere 21.2%, with problems such as vague and inconsistent data identified. In the meantime, the foundation's data website has experienced a notable surge in traffic driven by AI assistants. Currently, 26 UN-affiliated agencies have joined the system, with data from almost 20 agencies already available. The objective is to transfer 80% of the UN system’s statistical datasets by 2027. Google has allocated $2 million in funding for the platform’s development, which is independently operated and maintained by the UN. The platform supports data traceability and allows AI to automatically combine multiple indicators to create charts. However, conclusions generated by AI still require human oversight.