On February 28, at the inauguration and unveiling ceremony of the Artificial Intelligence Building, China Media Group (CMG) officially released the China Media Group Artificial Intelligence Development White Paper (2025 Edition). Guided by the concept of empowering media production and broadcasting with AI, leading media innovation, and reshaping the media ecosystem, the white paper outlines the overall architecture of CMG’s artificial intelligence application platform and sets forth CMG’s development path for generative AI technologies, providing guidance for accelerating the cultivation of new quality productive forces in the media sector.
In recent years, with the iterative upgrading of generative AI technologies, CMG has taken “CCTV Audio & Video Media Large Model” as its foundation and achieved coordinated progress in technological R&D and program production. This has led the innovation and upgrading of program formats and dissemination models, offering a new paradigm for comprehensively reshaping the product, technology, and application ecosystem of the media industry.
Generative AI
Empowering CMG Media Innovation
The newly released white paper provides a comprehensive review of the latest progress CMG has made in empowering media production with AI and leading media innovation, from the full production and broadcasting workflow of radio, television, and new media to the expansion of all-media content creation.
In terms of empowering the full production and broadcasting process, CMG has closely aligned with production needs and successfully developed and launched a series of AI applications, including intelligent subtitle generation, speech synthesis, intelligent review, intelligent horizontal-to-vertical conversion, and CMG algorithms, injecting intelligent capabilities into the production and dissemination chain and forming distinctive media AI applications that improve quality and efficiency at every stage of the production and broadcasting workflow.
In response to the rapid development of generative AI, CMG has deeply explored the leading role of AI large models in fostering innovative business formats and expanding application scenarios in the media industry. It has launched the “CCTV Audio & Video Media Large Model,” which demonstrates industry-leading technical capabilities in text-to-animation generation, sports tactics analysis, video creation, and editing.
Through the close integration of large-model technology and all-media production, CMG has launched a large-model toolset suitable for diverse program formats, accelerating the application of technologies such as digital humans, immersive AI virtual production, and AI radio dramas in the media industry, thereby driving innovation in communication formats and enhancing user audio-visual experiences.
“CCTV Audio & Video Media Large Model 2.0”
Reshaping CMG’s All-Media Production and Broadcasting Workflow
Looking ahead, the white paper focuses on CMG’s next-step plan to build an AI media application platform and develop the “CCTV Audio & Video Media Large Model 2.0,” presenting the overall vision, key systems, and implementation methods in the hope of offering reference and guidance for deepening the integration of media and AI and accelerating systemic transformation in mainstream media.
Building on its vast audio-visual corpus and rich program production and broadcasting scenarios, CMG will further strengthen the construction of its AI media application platform, accelerate the development of a multimodal Chinese corpus, and bring together several domestic large-model R&D institutions to form a joint R&D consortium. Together, they will create the industry-leading “CCTV Audio & Video Media Large Model 2.0” in the media sector, providing program production and broadcasting scenarios such as variety shows, sports, science and education, animation, and dubbing with media content creation assistance, personalized multimodal generation, and intelligent content creation capabilities, thereby building a more efficient and innovative intelligent media ecosystem.
The application of the “CCTV Audio & Video Media Large Model 2.0” will fully integrate large-model capabilities into all stages of all-media content planning, production, transmission, distribution, presentation, and evaluation. While empowering the entire all-media content production and broadcasting workflow, it will also simultaneously aggregate and process the massive audio and video data within CMG’s media resources and production systems, building a leading high-quality multimodal corpus in China and providing support for the development of China’s AI industry.
The deep integration of large-model technology will strongly promote system upgrading, workflow reengineering, and ecosystem reshaping in the media industry, accelerating the cultivation of new quality productive forces, new quality communication capabilities, and new quality competitiveness in the media sector, and driving the media industry toward a new stage of comprehensive and in-depth intelligent transformation and upgrading.
“Active exploration and application promotion,” “standardized use and safety assurance,” “technology leadership and quality improvement,” and “openness, inclusiveness, cooperation, and win-win” are the basic principles proposed in the white paper for CMG’s development and application of AI technology in the media field, and they also represent the “CMG experience” accumulated through years of exploration and practice.
The future has arrived, and CMG will continue to place innovation at the core of its competitiveness, offering more “CMG solutions” for the intelligent transformation of media worldwide.
