The piece titled “2023 Digital Human Practice Cases TOP10” is presented as a curated Top 10 list of “practice cases,” and its underlying publication format is explicitly case-based: for each entry it pairs a “user/implementer” (使用方) with a “technology provider” (技术提供方) and then names a specific project/case (案例), followed by a short narrative describing what was built, where it was deployed, and what practical value it was intended to deliver.
Across the ten entries, the selection emphasizes applied deployments in culture, media, commerce, and public-facing service rather than purely experimental demos. Case 1 frames a cultural-expo communications use: at the 19th China (Shenzhen) International Cultural Industries Fair, the digital human IP “Ayang” appears as a “digital explainer” in promotional short videos for China International Communications Group, with the write-up emphasizing real-time driving and real-time rendering to produce expressive facial and body animation that adds a “future-tech” feel to the booth presence and promotional output. Case 2 is a banking service case that centers on an “intelligent high-counter” system with an embedded digital human (“Xiaolan”) supporting a set of counter-like functions (voice interaction, face recognition, RMB exchange services, and remote virtual teller capability), and it notes an additional layer of model integration—after connecting to Baidu’s “Wenxin Yiyan,” the built-in virtual human is described as using data analysis of user needs to provide dialog, Q&A, operations support, customer service, and marketing functions, positioned as improving user experience in a “smart branch/outlet” context.
Several entries position “digital humans” as content engines and scalable presenters, highlighting AIGC-assisted production workflows. Case 4 describes a virtual digital host “Amanda” deployed at the 12th China Suzhou Cultural Expo as an on-booth guide/host for Suzhou Radio and Television Media Group, built with an AIGC video-generation tool (“魔珏有言”) to rapidly generate explainer videos with richer expressions and gestures, described as higher efficiency and lower cost than traditional video production. Case 6 similarly foregrounds a production pipeline: Kuaishangtong’s support for Gold Medal Kitchen Cabinet channel partners spans AI topic selection, AI scriptwriting, and AI video synthesis, where once a 2D ultra-realistic digital persona is built, the channel partner mainly performs text and video review, positioning the workflow as relieving capacity constraints in video making and “freeing” partners from labor-intensive recording processes. The e-commerce entry (Case 8) pushes this logic into sales operations: Baidu AI Cloud’s “Xiling” is described as a marketing large model trained specifically for e-commerce marketing scenarios, connectable to an end-to-end digital-human livestream workflow; the write-up claims that by combining Baidu’s foundation model plus an e-commerce domain knowledge base, a user can input keywords and have the system automatically generate product-introduction, urging-to-order, and live-room control scripts, then drive a digital human to “one-click” start a livestream, framed primarily as reducing labor cost in livestream selling.
Other entries highlight brand persona building and heritage/culture storytelling as the application core. Case 5 centers on Midea’s sub-brand Hualing and its digital brand lead “Linghun Shaonv,” described as a virtual brand figure used for consumer interaction; the narrative stresses a 2022 upgrade to a 3D digital human, alignment with Gen-Z aesthetics, and extension into lightweight consumer touchpoints such as a smart-connect mini program and NFC “stickers,” positioning the digital human as an organizing IP for digital marketing. Case 7 focuses on a historically grounded “hyper-realistic” digital Su Dongpo, described as jointly created by Zhonghua Book Company and Tingting Shijie using classical-document sources and digital technology, released around Zhonghua Book Company’s 110th anniversary and then featured on CCTV-1’s “Chinese Poetry Conference,” with the stated aim of combining “culture + technology + art” to create a new kind of stage experience bridging historical and contemporary contexts. Case 9 presents a regional cultural tourism “virtual star promoter” called “Qingniao,” first appearing at a Shanxi digital-culture branding event with technical support from Shanxi Zhisheng Media; the origin story ties the character to a Shanxi Museum artifact motif and to the mythic “Qingniao” as a messenger figure, and frames the project as transmitting Shanxi’s “thick/heavy” cultural heritage through a new digital mode and pushing it toward younger, trend-facing presentation.
Finally, the list includes two entries that foreground “digital humans” as productized systems that scale institutional output or explore new rights/business arrangements. Case 3 describes Xiaoice’s “cloned human” initiative in China and Japan and uses the “Hanzang Senlin” clone as the highlighted example inside the X Eva app: it is described as enabling real-time emotional interaction plus AIGC capabilities based on Xiaoice’s framework large model, neural rendering, and “super natural” voice, with a subscription payment model where proceeds can be transferred to the real person, explicitly framing it as an exploration of new digital-human applications and questions of rights/benefit attribution. Case 10 describes Xinhua Zhiyun’s “Yuanmao” digital-human family launched at an event jointly hosted by Xinhuanet and Xinhua Zhiyun, positioned as part of an AIGC-driven “metaverse system”; the digital humans are described as combining “real-person driving” and “intelligent driving” digital human technologies with computer graphics and AI to create ultra-realistic entities that can empower county-level integrated media centers for news production and experimentation with new distribution/communication modes, with the value claim focused on improving production efficiency and pushing new media production-and-spread models.
China International Communications Group (中国国际传播集团) - Shenzhen Cultural Expo digital explainer “Ayang” (阿央), a real-time driven digital presenter used in promotional short videos.
Postal Savings Bank of China (中国邮政储蓄银行) - “Xiaolan” (小蓝), a smart counter/virtual teller style digital-human service system for banking that adds voice, face recognition, and automated service functions.
Xiaoice (小冰公司) - AI “cloned girlfriend” Hanzang Senlin (半藏森林), a paid interactive companion-style digital persona tied to the company’s “cloned human” initiative.
Suzhou Radio and Television Media Group (苏州广电传媒集团) - Virtual digital host “Amanda” (Amanda) for the Suzhou Cultural Expo, produced via an AIGC video-generation tool for fast presenter-style content.
Midea (美的) - Hualing brand digital lead “Linghun Shaonv” (凌魂少女), a 3D digital persona used for youth-oriented brand interaction and marketing.
Gold Medal Kitchen Cabinet (金牌厨柜) - AIGC-enabled 2D ultra-realistic digital double video synthesis workflow provided by Kuaishangtong (快商通) for scalable channel content production.
Zhonghua Book Company (中华书局) - Hyper-realistic digital Su Dongpo (苏东坡) created with Tingting Shijie (谛听视界) and featured in CCTV’s “Chinese Poetry Conference.”
Baidu AI Cloud (百度智能云) - “Xiling” (曦灵), a marketing-focused model connected to end-to-end e-commerce digital-human livestreaming that auto-generates sales scripts and drives one-click broadcasts.
Shanxi Cultural Tourism (山西文旅) - Virtual “star promoter” digital human “Qingniao” (青鸟), a culture-themed persona derived from Shanxi Museum imagery to promote regional heritage digitally.
Xinhua Zhiyun (新华智云) - “Yuanmao” (元卯) digital human family, an AIGC-driven set of ultra-realistic digital humans for county-level media news production and new dissemination modes.