Tashkent, Uzbekistan (UzDaily.uz) — Xinhua News Agency presented its AI agent, Xinhua Yudian, at the 7th World Media Summit in Beijing. According to the agency, the platform is powered by verified Xinhua content and helps users search for, verify and use news information.
The platform was launched on 26 June 2026, according to a Xinhua representative speaking at the summit.
Summit materials describe Xinhua Yudian as an authoritative AI assistant for current affairs, built on what the agency calls a “clean and reliable” data corpus. The platform combines four service modules: an AI chat, subscriptions and notifications, a knowledge base, and an AI assistant hub.
Users have access to seven main functions: studying theory, searching for primary sources, receiving news recommendations, asking questions about news, verifying information, analysing topics and preparing articles. According to Xinhua, the service is designed to provide authoritative, reliable, secure, accurate and traceable information throughout the AI process. It aims to move beyond the basic delivery of content from mainstream media towards in-depth knowledge and services.
The representative said the project was developed in response to the growing number of people relying on AI to obtain information and the widespread use of intelligent tools by media organisations for research, content creation and information gathering.
She noted that this trend raises questions about the accuracy of AI-generated answers, whether their sources can be verified, and whether systems correctly understand and cite professional news reports. She described these as matters of public trust and the future of professional journalism.
According to Xinhua, the service draws on the agency's 95 years of reporting experience. To cover major statements by Chinese President Xi Jinping, the system incorporates more than 100 volumes of documents and hundreds of thousands of full-length news reports. The agency has developed specialised knowledge graphs and networks of entity relationships intended to ensure accurate understanding, retrieval and citation of content.
On the technological side, the Xinhua representative said the agency independently develops the platform's core architecture, fundamental algorithms and main modules. To address AI hallucinations and improve information traceability, the system combines multi-channel knowledge-base retrieval, knowledge graphs and deep semantic search. Security is provided through a three-tier protection system covering the entire knowledge infrastructure and AI application process. The agency has also taken the lead in promoting standards in related industries.
Xinhua said senior experts in cognitive technologies at the Chinese Academy of Sciences view the project as an effort to develop highly reliable content-based AI, with potential demonstration value for authoritative information services provided by mainstream media.
According to the agency, Xinhua Yudian is already being used in practice for public opinion research, policy studies and assistance with drafting texts.
Xinhua has invited media organisations from other countries to pursue practical cooperation. The agency is ready to share its experience in building specialised text corpora and jointly discuss ways to organise and use historical reporting, research materials and professional knowledge more effectively. It also seeks cooperation on multilingual and intercultural services that take into account the history, culture and communication practices of different countries.
Earlier at the summit, Xinhua presented its Global Media Initiative to jointly develop standards for AI use, while participants adopted the Beijing Consensus.
