← J·Index — The Journalism AI Index
J·Index / China / China Central Television (CCTV)

China Central Television (CCTV)

3 verified, source-linked AI implementations documented between 2019–2025.

ChinaAsia-PacificNational
Cases
3
Active since
2019
Peak stage
4
Controversies
0

Where the AI sits

Editorial Production3

AI implementations

2025

央视听媒体大模型 v2.0

Upgraded large model trained on high-quality variety shows, dramas, animation and sports content; focuses on newsroom production scenarios and multimodal collaborative smart media applications; launched February 2025

Editorial Workflow & Writing Assistance4 – Ways of Working3 – Collaborative Human–AI
Tool: 央视听媒体大模型 v2.0Source ↗
2023

央视听媒体大模型 (CCTV Media Large Model) v1.0

Joint development with Shanghai AI Lab; multimodal AI model trained on CCTV audiovisual archive; capabilities include intelligent editing, scene identification, narration generation, digital anchor creation from brief real-person samples, and sports highlights editing (Olympics coverage)

Multimedia & Visual Production3 – Proficiency3 – Collaborative Human–AI
Tool: 央视听媒体大模型 v1.0 / Shanghai AI LabSource ↗
2019

AI virtual news anchors (Kang Xiaohui, Xiaoling, Tongtong, Ji Xiaomeng, Xiaozhi)

Synthetic presenters for varied broadcasts; Xiaoling used specifically for CCTV's May Fourth Party broadcast; developed with Faceunity, Iflytek and Sogou

Multimedia & Visual Production3 – Proficiency3 – Collaborative Human–AI
Tool: Faceunity Technology; Iflytek; SogouSource ↗

Context

China Central Television (CCTV) is one of 58 organizations in China documented in the index, which holds 75 cases from the country. Its work spans 2 function types, including Editorial Workflow & Writing Assistance, documented across 283 organizations worldwide.

Explore related

More newsrooms in ChinaEditorial Workflow & Writing AssistanceMultimedia & Visual ProductionHow this is verified
Part of J·Index — The Journalism AI Index. Every case carries a live source link. Cite with attribution; bulk reproduction is not licensed. Data licensed CC BY-NC 4.0.