CCTV International Network Co., Ltd. (央视国际网络有限公司) has been actively developing and deploying “digital human” technologies across multiple domains, with applications in broadcasting, live streaming, cultural promotion, education, and AI-powered media production. The company has filed patents such as a method for a digital human to automatically narrate static PPT presentations to improve video playback flow, and has developed 3D ultra-realistic digital humans for complex live scenarios, including its award-winning virtual presenter “Xiao C.” It participates in industry initiatives like the China Digital Human Pilot Program and has showcased work at events including the AI Digital Human Forum and the China Network Audiovisual Conference. Its digital human capabilities are integrated with its “Zhice” large media model and other AI platforms to enhance media fusion, automated content creation, and multilingual delivery. Projects span realistic, cartoon, and virtual character types for entertainment, government services, tourism, and e-commerce, with examples including municipal service avatars, cultural ambassadors, and digital human e-commerce hosts.
CCTV International Network Co., Ltd. is a wholly owned subsidiary of China Media Group (CMG), established on April 28, 2006, as the group’s core platform for online new media operations. It operates CCTV.com, the central key news website and national online audiovisual broadcasting organization, with full licensing qualifications for multi-terminal business. Its origins trace back to December 1996, when the CCTV website launched as one of China’s earliest Chinese-language news portals. The company integrates television content with internet features, focusing on news, video, and user engagement, and holds exclusive rights to distribute CCTV’s television programming online. Headquartered in Haidian District, Beijing, it has more than 1,000 employees and manages multiple sub-brands, including China Network Television (CNTV, 2009–2016). It undertakes the mission of promoting Chinese culture globally, implements the CMG strategy of “equal emphasis on TV and online platforms, online first, TV support,” and is active in intellectual property protection, digital media innovation, e-commerce, live streaming, and international broadcasting.
CCTV has played a prominent role in showcasing and adopting digital humans across its programming, using AI-generated avatars in news broadcasts, special events, and public information campaigns. These digital figures serve as virtual anchors, sign language interpreters, brand representatives, and even recreations of historical personalities. CCTV has collaborated with major tech companies to implement technologies such as real-time motion capture, facial cloning, large language models, and speech synthesis. The network emphasizes the growing role of digital humans in media, governance, education, and commerce, while also addressing concerns about ethical standards, realism, legal frameworks, and professional development. Through frequent coverage and deployment, CCTV positions digital humans as a core element in China's digital transformation strategy.
“CCTV Listen Media Model 2.0” (央视听媒体大模型2.0) is described as a self-developed large-model system from China Media Group positioned as an AI toolchain for end-to-end audiovisual production in which digital humans are treated as a core output format rather than an add-on: news reports note it was first applied to CCTV Spring Festival Gala content creation, where it is used for generating image/video materials and producing “hyper-realistic” digital humans to improve production quality and efficiency, and it is also presented as a production enabler in major-event live broadcasting, including 15th National Games of China coverage where the model supports functions such as “video intelligent creation,” automatic highlight clipping, intelligent planning, and Chinese–English–Cantonese intelligent translation, alongside “light-field scanning” hyper-realistic digital humans and “lossless” digital-human modeling used to build immersive interview scenes with mixed virtual/real presentation; the same news reports link the product to CMG’s broader “5G + 4K/8K + AI” workflow framing and to a 2025 white-paper discussion that associates the model with lightweight 3D modeling and virtual production pipelines (e.g., LED virtual production and XR interactive applications), and one listed item further characterizes the 2.0 version as having completed a 4K hyper-realistic digital-human upgrade in February 2025 with multilingual simultaneous broadcast capability that was applied in “Two Sessions” reporting via a virtual anchor workflow.