February 26 News
Public Services And Government: In Xi’an, a 3D digital human smart customer service agent named Xiaolian was deployed in a public service hall to answer citizens’ questions by voice, functioning as an on-site intelligent service interface in a government-facing administrative environment rather than a pure text chatbot. In Beijing’s digital economy computing center, a customized government AI digital human was reportedly built for Beigou Village in Huairou, indicating a localized public-service deployment model in which a screen-based virtual agent is tailored to a specific community’s administrative needs. In Hubei’s Xiaogan municipal service reform pilot, authorities also described a standardized smart government service matrix that includes AI customer service and government digital humans designed for second-level precision response, showing a broader pattern of digital-human integration into public-service delivery workflows.
Culture And Tourism: Chengdu’s Du Fu Thatched Cottage Museum presented a concrete cultural heritage case in which visitors interacted with a digital human version of Du Fu and engaged in poem exchanges with Li-Du themed digital/NPC characters during the Spring Festival period, combining literary heritage interpretation with interactive virtual character experiences in a museum setting. Related reports on Chengdu holiday tourism and Qingyang District consumption activity confirm that the Du Fu digital human became a visible attraction within a wider technology-enhanced tourism environment that also included other smart installations, demonstrating how digital humans are being used as experiential cultural interpreters rather than only as display media. In Shanghai’s international business cooperation zone pilot area, digital humans and LED screens were used as “intelligent guides” in a controlled transit and customs-service context, extending virtual-being use into wayfinding and service support for cross-border mobility scenarios.
Education And Learning Platforms: Guangdong’s provincial lifelong learning platform introduced a notable educational-cultural application through historical-figure digital humans, with an initial group including Ge Hong and Zhang Jiuling and other figures linked to Guangdong, allowing users to “converse” across time with historically themed virtual characters. This positions digital humans as an interface for public learning, cultural education, and engagement with local history on a province-level digital education platform rather than as isolated promotional mascots. The implementation is notable because it embeds named historical digital humans into a formal lifelong-learning system, tying virtual character interaction to institutional learning infrastructure and credit-linked educational services.
Enterprise Operations And Workplace Systems: Yunnan Nantian Electronic Information Industry Co., Ltd. (云南南天电子信息产业股份有限公司) registered software for a Smart Office Digital Human Platform V2.0, indicating a named enterprise software product oriented toward office scenarios and suggesting continued productization of digital humans for internal workplace and administrative functions. In Henan, Foxconn Zhengzhou (郑州富士康) was reported to use an AI-built virtual human named Xiaoqing in its “lighthouse factory” environment, where the virtual human can retrieve required information from very large data volumes and support rapid analytical output, showing a manufacturing intelligence use case in which a virtual human acts as an information-access and decision-support interface. These examples together show two distinct enterprise directions: office-platform software for digital-human deployment and factory-floor data interaction through a named virtual human assistant embedded in industrial operations.
Commerce, Marketing, And Livestreaming: Baidu, Inc. (百度集团) reported large-scale commercial use of merchant intelligent agents, alongside digital-human lead-generation livestreaming that was said to improve merchant conversion performance, and one cited case involved a Beijing vocational education institution adopting a voice-enabled intelligent agent version, demonstrating an integrated sales and customer acquisition workflow that combines conversational AI with digital human presentation. A separate report on a company deploying 94 “digital employees” described livestreaming operations in which digital humans respond to bullet-screen comments and synchronize reply content with a host’s spoken script, as well as switching scripts by product link, which is a concrete example of digital-human use in high-volume ecommerce content operations and semi-automated customer engagement. Reports also referenced AI digital human commercial development in cross-border ecommerce operations, indicating that digital-human capabilities are being bundled with livestream planning, storefront operations, supply-chain coordination, and after-sales support in platform commerce contexts.
Media Production And Broadcasting: Chinese media-related reports described super-realistic digital humans being used as virtual commentators in large-event content production pipelines, combined with ultra-high-definition multi-camera capture, AI-assisted tracking, and immersive audio technologies, indicating a production model in which synthetic characters are part of a broader broadcast technology stack. Another Beijing-related item noted that a video was produced using Huiboxing Digital Human, which reflects the use of a named digital-human production tool in public-facing media content. These cases show digital humans functioning as presentational and explanatory personas in media production, especially where organizations need scalable on-screen hosts or commentators without relying exclusively on conventional human presenters.
Consumer Social And Creator Technology: Soul App open-sourced SoulX-FlashHead, a real-time digital human generation model positioned for lightweight, consumer-grade real-time high-fidelity livestreaming, with reporting emphasizing a technical tradeoff it attempts to solve between image quality and computational cost. This is a concrete product and technology case because it names the platform, the model, the deployment context (real-time livestreaming), and the intended accessibility target (consumer-grade hardware), making it relevant as an enabling technology for synthetic character creation and live operation. The open-source release also indicates a shift from closed proprietary pipelines toward developer-facing infrastructure for real-time digital human production in China’s social and creator ecosystem.
Identity, Family Memory, And Personal Virtualization: Mever (元我宙Mever) launched the Mever Space in Zhangjiang Science City, Shanghai, with its founder describing a model centered on creating a personal virtual digital human for each user, linked to long-term recording and inheritance-oriented uses such as moving from diary-like records toward family lineage preservation. This is a concrete virtual-being case because the product framing is explicitly tied to individualized virtual digital humans rather than generic AI tools, and the deployment context is a named platform launch. The reported positioning suggests a distinct application category in China’s virtual-being landscape focused on digital identity continuity, personal archiving, and intergenerational memory representation through persistent synthetic characters.
Mever Space (元我宙Mever空间) is positioned as a digital-life platform developed by Yuanwozhou (Shanghai) Information Technology Co., Ltd. (元我宙（上海）信息技术有限公司) that uses digital humans as a personal “AI digital double” layer on top of long-term memory capture and inheritance workflows, with the stated aim of turning everyday records into a persistent virtual human that can reflect a user’s thinking patterns and emotional memories. In this framing, the digital human is not a standalone character product but an identity-linked synthetic persona generated from user behavior and personal media, designed to operate inside a broader system that supports multi-format journaling and biographical records, family co-creation spaces for shared albums and chronologies, and an AI-assisted genealogy structure that models kinship relationships over time. Its technical description emphasizes an underlying graph-style relationship model connecting people, families, events, and timelines, deep learning for generating the high-fidelity digital double, blockchain for durable and tamper-resistant storage, and cross-platform access through WeChat, app, and web interfaces, placing the digital human within a platform architecture oriented toward personal archiving, family history continuity, and intergenerational memory representation rather than short-term customer service or entertainment hosting.
February 25 News
Healthcare: Changzhou Hospital of Traditional Chinese Medicine in Jiangsu deployed the traditional Chinese medicine digital human “Dr. Meng” as a clinical-facing service example tied to local “smart doctor” rollout efforts. The system combines clear voice-based inquiry with tongue-surface AI recognition, pulse-sensor data collection, interactive questioning, and automated generation of personalized diagnosis-and-treatment plans, and it was built through a joint collaboration involving Changzhou Hospital of Traditional Chinese Medicine, Nanjing University of Chinese Medicine, China Mobile Changzhou Branch (中国移动常州分公司), and Nanjing Silicon Intelligence Technology Co., Ltd. (南京硅基智能科技有限公司), showing a hospital deployment context in which a digital human is integrated with multimodal sensing rather than used as a standalone chatbot. A separate parenting-use case involved a doctor’s AI double made available through the Afu App, where new parents used the virtualized doctor persona to ask questions during the Spring Festival period because they trusted the original physician, indicating an AI companion or digital clone model used for anxiety reduction and family health support in consumer care scenarios.
Public Services and Utilities: State Grid Hubei Information and Communication Company (国网湖北信通公司) deployed AI “employees” in power service halls in Wuhan, including digital human interfaces used in physical electricity business halls during the holiday period. These digital staff members did not function only as passive Q&A tools, because they also proactively pushed time-of-use electricity price comparison visuals and explained savings logic such as cheaper nighttime charging, which positions the digital human as a service counter agent embedded in a utility workflow with advisory and guidance functions. In a separate public-facing government and infrastructure setting, Beijing’s “super-live” ecosystem push described digital human cloning as part of multimodal interaction design, where real-time question answering and digital avatar interaction are used to extend live event experiences beyond conventional viewing and to support interactive participation.
Transport, Airports, and Tourism Access: Shanghai’s Eastern Hub International Business Cooperation Zone startup area used digital humans alongside LED-based guidance systems as “intelligent guides” to support emergency response and smoother customs and clearance services, making the virtual-being component part of border-adjacent business travel operations rather than an entertainment feature. In Hunan, Changsha Airport introduced digital human guided-tour functions that let travelers experience a virtual tourism-style presentation of Hunan cultural destinations while physically inside the airport, demonstrating a deployment model that blends passenger service, regional promotion, and travel experience enhancement. In Hangzhou-related aviation technology showcases, “virtual digital human” installations were listed together with AI image generation and other interactive experience points, indicating a themed airport or travel-tech exhibition context where virtual beings were presented as part of an integrated public technology experience.
Culture, Museums, and Education: Zhejiang museum programming for the Spring Festival period included a chat-capable “Dongpo digital human,” showing a heritage interpretation use case in which a historical or cultural figure is rendered as an interactive virtual being for visitor engagement. This is notable because the function described is not merely display animation but conversational interaction, which places the application in the category of synthetic cultural characters used for public education and museum mediation. In Hunan education discourse, “AI digital humans” were also described as interactive participants in ideological and political education, with the stated aim of increasing immersion and student engagement through direct exchanges between digital human figures and students.
Entertainment and Performance Media: China Media Group (中央广播电视总台) used a hyper-realistic digital human technique in the Spring Festival Gala to create a digital double of the “dream figure” in a performance sequence, presenting multiple versions of the same figure on screen simultaneously to construct a dream-state narrative effect. This is a concrete production case of digital human compositing in a flagship broadcast setting, with the virtual being used as a narrative and visual performance device rather than as a service agent. Hunan Broadcasting System (湖南广播影视集团) was also described as staging a performance in which singer Liu Yuning appeared with an AR digital human, indicating live-stage integration of augmented-reality virtual beings in televised entertainment. Separate reports on the program “Ride the Wind 2026” described a virtual idol dance appearance that triggered audience backlash over movement quality and perceived “heat-chasing,” with discussion centering on poor motion-capture execution and technical debugging problems, which still constitutes a concrete case study of virtual idol deployment and failure in a mainstream variety-show context. Commentary around Cai Ming additionally referenced her established virtual UP host persona “Caicaizi Nanako,” showing how a human celebrity can maintain a parallel synthetic character identity in a different audience ecosystem.
Enterprise, Workplace, and Commercial Operations: Baidu, Inc. (百度) circulated a digital human version of founder Robin Li to employees on the first workday after the holiday, using the Ruliu workplace platform service account to distribute reopening “red envelope” messages. This is a clear internal enterprise communication use case in which a named executive’s digital double is operationalized inside a corporate collaboration environment, demonstrating digital-human deployment beyond marketing and customer service. In procurement and tendering scenarios, ZeYu AI (泽宇AI) was described in relation to bidding cases that introduced digital human presentation-and-defense processes to assist evaluators in comparing shortlisted candidates and making final decisions, indicating use of virtual human interfaces in formal evaluation workflows. Guangdong examples also pointed to Quwan Technology (趣丸科技) using digital human capabilities in conjunction with audio interaction and voice-model technologies, including a digital volunteer application for the 15th National Games, which shows event-service deployment of virtual beings in a large public event context.
February 24 News
Healthcare: A Hong Kong Wen Wei Po report described a care model centered on collaboration between digital humans and human caregivers for an aging population, presented by Feng Hui of the Xiangya Pan-Health Management Research Institute at Central South University. The reported emphasis was not a generic AI assistant but a “digital human and human-machine collaborative” care framework for eldercare, indicating a deployment context in senior health management and daily care coordination where digital-human interfaces are positioned as part of service delivery.
Culture and Tourism: A Spring Festival report circulated through the Tibet regional cyberspace authority described a concrete consumer-facing use case in which a 63-year-old resident from Hangzhou, He Xia, published a New Year greeting video whose on-screen protagonist was an AI-generated digital human, showing personal festive media creation as an actual deployment scenario rather than a conceptual demonstration. In a separate museum case, the Du Fu Thatched Cottage Museum in Chengdu was reported as offering visitor interaction with a digital human version of Du Fu and an on-site NPC-based Li Du poetry exchange experience, placing digital human characters in a cultural heritage setting for interpretation, engagement, and interactive literary participation.
Retail and Consumer Experiences: Alibaba (阿里巴巴) was reported as using virtual digital human customization inside the “Milan Ali Gallery” exhibition to support large-scale immersive participation and AI-assisted shopping experiences. The described workflow combined head-mounted displays, visual recognition of a visitor’s clothing style, and question-and-answer interaction to generate a personalized virtual avatar, then linked that avatar to try-on functions across themed outfits, which indicates a multi-step retail pipeline integrating identity generation, styling interaction, and immersive commerce presentation in a single deployment.
Media and Entertainment: Volcano Engine (火山引擎) was reported as enabling a Spring Festival consumer experience in which users upload a selfie to generate a virtual avatar that can perform in AI-enhanced skits alongside celebrity AI doubles, including motion-capture-enabled effects workflows packaged for ordinary users rather than only production studios. Taiwan Mobile (台灣大哥大) was also referenced through its MyMoji feature, where AI-avatar interaction functions were upgraded for holiday gameplay, indicating continued commercialization of AI “self” representations in telecom-linked consumer entertainment services. In a separate entertainment case, reports about the virtual idol Asuka described a company-announced storyline involving marriage to an AI-generated character and a planned “cross-dimensional marriage” product direction, which frames the AI-generated counterpart as a synthetic character within an idol fandom and monetized virtual relationship context.
Public Service and Civic Contexts: Beidian Shuzhi (北电数智) was cited in reporting on Beijing regional innovation activity for an “AI digital human village secretary” located in the Jiuxianqiao area’s Beijing Digital Economy Computing Center, indicating a public-service or governance-oriented digital human use case rather than a purely commercial avatar application. This case places a named digital human role inside a civic administration context and links it to local digital infrastructure development, suggesting deployment around public communication, service guidance, or community-facing governance interaction.
Enterprise and Financial Services: Hunan Bank (湖南银行) appeared in procurement-related reporting through a 2025 digital human system construction project and related tender notices, showing that a bank-level digital human initiative had progressed into formal project bidding and candidate selection. Although the reports do not provide the final implementation architecture or vendor-delivered feature set, the documented procurement stage confirms an enterprise deployment context in which a dedicated digital human system was planned as a defined banking technology project rather than a general AI discussion.
News and Content Production Workflows: Multiple reports carried production tags indicating use of the Huiboxing Digital Human (慧播星数字人) for video generation, including local-news-style clips distributed on mainstream portals. This is a concrete operational use case for a named digital human production tool in media publishing workflows, where the digital human functions as an on-screen presenter format for packaged news or commentary videos rather than a standalone chatbot or voicebot.
February 23 News
Marketing: Baidu (百度) is referenced through its Huiboxing digital-human presenter, which is used to deliver short-form video segments that frame commercial and public-interest topics as on-camera “digital human” narration, and is also described as a live-commerce capability that can be opened to merchants as a standardized digital-human selling workflow. JD.com (京东集团) is described as making its official digital-human livestreaming capability available to all merchants on its platform at no cost, with access routed through the JD Jingmai (京麦) service marketplace, positioning digital humans as a scalable front-end for merchant-hosted sales sessions rather than a one-off brand activation. Jiufang Zhitou (九方智投) is presented as offering a “Jiufang digital human” alongside AI functions used in finance-oriented workflows, indicating a productized digital-human interface aligned with analytical and content-generation tooling that can be packaged into investment-information services and related customer-facing experiences.
Jiufang Zhitou is a China-based securities investment advisory company that has positioned digital humans as a user-facing interface for delivering AI-assisted investment guidance inside its app, most notably through an intelligent investment-advisor digital human commonly called “Jiugo” (九哥). In this framing, the digital human functions as the conversational, personified front end for market interpretation and stock-related explanations, while the underlying capability is presented as a broader “AI + investment research” stack that includes dialogue and research modules branded as Jiufang Lingxi (九方灵犀) and Jiufang Zhiyan (九方智研). Publicly reported usage figures associated with this digital-human advisor have been described at scale across 2024 (hundreds of thousands of clients and tens of millions of cumulative service interactions), and descriptions of the experience emphasize human-like audiovisual presence and natural-language interaction as a way to make financial analysis and advisory workflows more accessible and continuously available to retail investors.
Cultural: Chengdu Du Fu Thatched Cottage Museum is described as deploying interactive digital-human versions of Du Fu and Li Bai for on-site Spring Festival programming, where visitors conduct real-time dialogue with the figures and participate in poetry exchanges that are framed as interaction with both a digital human and an NPC counterpart in the same setting, using the museum environment and associated scenic nodes as the staged interaction context. A separate Beijing Haidian temple-fair setting is described as featuring a multilingual digital human modeled on the Yuanmingyuan bronze animal-head motif (the horse head), with the character switching across multiple languages and responding extemporaneously to audience questions, indicating an emphasis on live, public-facing cultural interpretation and tourist engagement through an embodied synthetic character rather than static signage or pre-recorded audio.
Public Communication: NetEase (网易) and Sina Finance are both described as using Huiboxing’s digital-human presenter format for public-information videos, including incident narration and rumor-debunking clips, which places a digital human in the role of a standardized anchor or explainer persona rather than a bespoke character created for a single campaign. In these examples, the digital human functions as the delivery layer for safety messaging and AI-misinformation rebuttal, showing the deployment context as rapid-turnaround media publishing where the virtual being is used to present statements, warnings, and clarifications in a consistent on-screen format.
February 22 News
Public Sector And Urban Services: A multi-ministry national policy package led by the National Development and Reform Commission and the Ministry of Industry and Information Technology explicitly called for introducing digital humans into public-service and governance modernization efforts, positioning “digital humans” as a deployable interface for service delivery alongside broader digital transformation measures. In holiday travel and city-facing service scenarios, AI digital humans were described as “navigation” helpers for travelers during the Spring Festival return-journey context, framed as a practical, front-end guidance layer integrated into public-facing convenience services rather than a back-office chatbot.
Transportation, Retail Spaces, And On-Site Experiences: In travel and retail-footfall environments, digital humans and avatar-like characters were presented as on-site engagement and guidance components, including “AI digital human navigation” for travelers and the use of a “Little Horse” digital human derived from the Yuanmingyuan bronze horse-head cultural motif as a themed interactive presence in a public venue context. In Beijing’s Haidian Spring Festival science-and-technology fair setting, digital humans appeared as part of an interactive portfolio that also included large-model experiences and VR/AIGC demonstrations, indicating that digital humans are being deployed as participatory, visitor-facing interfaces within municipal showcase events rather than as standalone products.
Entertainment, Broadcast Production, And Virtual Performance: Spring Festival Gala-related reporting emphasized virtual hosts and AI-presented segments as a visible application of virtual beings in broadcast packaging, alongside XR virtual-stage production that blended performers with stylized virtual environments to create tightly integrated “virtual–real” staging. ByteDance (字节跳动) was described as participating in multiple Gala production elements through the Doubao model family and Volcano Engine (火山引擎), including workflows that transform a real performer into a 3D “digital double” for stage use and that tune interaction details such as voice and speaking style for human–machine performance beats. Lingyun Guang (凌云光) was described in connection with the Gala stage program “Meng Di” (梦底), where performer Liu Haocun appeared with five sub-millimeter, high-precision digital doubles on the same stage; the account attributed this to high-density multi-view synchronized capture and light-field information modeling that supported cross-time-stage interaction effects and coordinated choreography with multiple virtual replicas.
Marketing, E-Commerce, And Brand Promotion: Alibaba Group (阿里巴巴集团) was described as using digital-human “colleagues” in Taotian Live to cover late-night operational periods, framing digital humans as labor-substituting on-camera or hosting capacity within livestream commerce workflows rather than merely as editing tools. BlueFocus (蓝色光标) was cited in an investment-themed discussion as part of an “AI content/marketing” narrative tied to Doubao-related virtual-hosting and AI marketing within the Volcano Engine ecosystem, pointing to agency-side packaging of virtual hosts as scalable campaign assets. In regional B2B marketing-services positioning, Gansu New Generation Network Technology Co., Ltd. (甘肃新一代网络科技有限公司) and Tianjin Hanfan Digital Technology Co., Ltd. (天津汉梵数字科技有限公司) were described as offering “AI digital human anchors/hosts” and related marketing toolchains (including matrix-style account systems and broader AI content production capabilities), presenting digital humans as productized components inside bundled customer-acquisition and content-automation offerings.
Culture, Tourism, And Heritage Communication: In a cultural-heritage venue upgrade, a globally premiered “Li Bai digital human” was described as a cultural communication envoy within the Li Bai Memorial Hall context, paired with AR light-and-shadow shows and holographic projection that recreated a poetic “walking chant” depiction; this positioned the digital human as a narrative guide and symbolic mediator connecting historical figures to contemporary audiences. (The published bid result lists a winning consortium of Beijing Baige Era Culture Development Co., Ltd. (北京百舸时代文化发展有限公司) and Beijing Fancai Digital Technology Co., Ltd. (北京凡彩数字科技有限公司).) In city-and-scenic-area consumption contexts, “scenic-area digital humans,” metaverse exhibition halls, and digital light-and-shadow exhibitions were described as part of Spring Festival tourism consumption experiences, indicating that digital humans are being deployed as interpretive hosts and interactive presences inside location-based attractions rather than solely as online media characters.
Enterprise Platforms, Capability Assessment, And Named Digital-Human Applications: CapCloud (开普云) was reported as having its AI digital human capability platform pass a digital-human system capability assessment associated with a national-level industry institute evaluation process, framing digital humans as an enterprise platform category with testable system competencies rather than only a content format. In a brand-ambassador deployment example, Huayi Zhiguang (华翼智光) launched a public beta for its brand digital human “Moliya” (摩力娅) 1.0, positioning it explicitly as a named branded virtual being intended to represent and front a product or organization identity in a controlled rollout.
Ethical Contexts, Misuse, And Misinformation Workflows: Several items described the use of a named digital-human presenter tool, Huiboxing digital human (慧播星数字人), as the production layer used to render presenter-style videos in politically themed commentary clips and in rumor-debunking contexts, illustrating how digital-human anchors can be used as low-friction synthetic spokespeople regardless of the underlying claim’s quality. Separate reporting described “AI digital human” video content that targeted older viewers by addressing them directly and prompting them to leave contact details, with the concern framed around privacy capture and manipulative lead-generation tactics; in this depiction, the virtual being is the persuasive interface, while the risk lies in downstream data extraction and coordinated “surround” marketing. These cases collectively situate digital humans as a credibility-bearing presentation wrapper that can be used for legitimate narration, but also for deceptive or coercive flows when coupled with fabricated claims or personal-data harvesting.
February 21 News
Entertainment: China Media Group (中央广播电视总台) integrated hyper-realistic digital humans into the 2026 CCTV Spring Festival Gala as part of an “AI Spring Festival Gala” framing that combined stage performance with generative production workflows, including digital-human segments described as being generated by an in-house large model and presented as dreamlike or immersive vignettes rather than as conventional on-screen hosts. ByteDance (字节跳动) was positioned as a key enabling vendor through Volcano Engine (火山引擎), described as supporting large-scale real-time interaction, while the Doubao model was described as being used to deliver mass-scale “tech gift” style interactions into households, tying a consumer-facing distribution moment to model-driven content and digital-human presentation. In parallel, Wenzhou’s Longwan District publicized a winning entry from the “Longwan Light” digital idol competition as a stand-up style program titled “AI’s Spring Festival Anxiety,” in which a virtual being is shown traveling through specific Longwan landmarks in Wenzhou (including Wenzhou Olympic Sports Center, Zhongxiu Garden, Qianfo Pagoda, and Yunye Life), using an AI camera perspective to blend comedic performance with location-based virtual presence.
Cultural: Multiple city and venue cases framed digital humans and avatars as a layer added onto New Year cultural programming, with deployments emphasizing guided storytelling, commemorative interaction, and place-based “new year tech” experiences rather than generic conversational service. In Xi’an, the “Xi’an New Year, Most China” digital culture-and-tourism program (“Digital Xi’an New Year”) explicitly positioned digital humans as a way to narrate cultural relic stories more accessibly to the public, treating the virtual being as a mediator between artifacts and visitors within a festival tourism context. In Guangzhou, consumer-facing New Year greetings were described as being produced by animating people in old photos so that the depicted individuals appear to speak, then pairing that effect with “nimble” digital-human presentation to perform personal messages, aligning memory-triggered media with a virtual-being delivery format. These cultural deployments were further complemented by a campus-linked announcement of HuaYiZhiguang (华翼智光) launching a branded digital human named Moliya 1.0 as a public beta, explicitly framed as a “brand ambassador” type virtual being and presented as a named, versioned character rather than an anonymous animation effect.
Public Services And Transport: Several reports framed digital humans as frontline guidance and service intermediaries during Spring Festival travel peaks, emphasizing wayfinding, query handling, and multilingual holiday messaging in high-footfall environments. A Shaanxi travel-technology report described an “AI digital human” at a service-area fueling/charging station functioning as an all-purpose guide for travelers, handling tasks such as navigation queries and charging-related consultation in a deployment context oriented around road-trip support and transport infrastructure. Separately, a broadcast news item credited the use of a HuiboXing digital human (慧播星数字人) to produce or present video coverage related to a Hunan service-area assistance incident, reflecting a newsroom-style use of a named digital-human product to deliver public-interest reporting content in a standardized presenter format. In addition, public-facing festival environments were described as featuring “little horse” digital-human style holiday greetings and other digital-human appearances alongside broader technology exhibits, with the digital-human layer specifically used for greetings and guided public interaction rather than for physical task execution.
Enterprise And Marketing: Chinese vendors were presented as pushing virtual-being production toward rapid, scalable commercial media generation, with an emphasis on “AI doubles,” brand spokes-characters, and short-form video pipelines that can be aimed either at consumers or enterprise clients. ByteDance (字节跳动) was highlighted in this commercial framing through references to “AI doubles” used by online creators and through the naming of Seedance 2.0 as a capability associated with fast generation of cinematic-style video, positioning the virtual-being layer as part of an end-to-end content-production stack rather than a standalone character. Kuaishou Technology (快手科技) appeared in the same commercialization context via the positioning of Kling as “to B,” implying enterprise packaging for AI video generation that can incorporate virtual presence, avatar-like transformations, or creator-facing “AI double” workflows for marketing content. Alongside these platform-scale pipelines, HuaYiZhiguang (华翼智光) treated its named digital human Moliya 1.0 as a brand asset with a defined release state (public beta) and explicit ambassador role, aligning the virtual being with productized brand identity rather than one-off campaign animation.
February 20 News
Cultural: China Media Group (中央广播电视总台) deployed a digital human representation of Liu Sanjie as a named synthetic character in a Guangxi New Year stage production located in Nanning, using the virtual performance to anchor regional cultural identity and to blend singing with landscape-themed visual staging that referenced Guilin scenery. In Sichuan, the Du Fu Thatched Cottage Museum in Chengdu presented an interactive AI digital human version of Du Fu as a visitor-facing avatar, positioned as an on-site engagement layer for Spring Festival tourism crowds and designed for direct public interaction inside the museum environment. In Shaanxi, Xi’an Museum used digital human presentation to make cultural relic narratives more accessible, explicitly framing the digital human as a mechanism for telling deeper artifact stories to broaden public understanding of heritage objects rather than treating the virtual figure as a purely decorative effect.
Entertainment: China Media Group (中央广播电视总台) and the technology teams supporting its Spring Festival Gala programming integrated hyper-realistic digital human techniques, XR stage extension, and AIGC-based spatial augmentation to place virtual beings and human performers into the same staged sequences, including scenarios where digital doubles and synthetic characters appeared across shifting time-and-space compositions to intensify narrative and emotional beats. ByteDance (字节跳动) positioned its Doubao virtual-being marketing around a real-time virtual-human interaction system built for gala-scale participation, with Volcano Engine (火山引擎) described as providing the technical team that implemented the live interaction stack, including the cost-intensive setup work needed to run real-time virtual-human experiences under broadcast constraints. In parallel entertainment discourse around variety programming, Mango TV’s Ride the Wind 2026 concept was discussed as staging competition formats that place virtual idols alongside human performers, with prior audience reaction cited for the virtual guest Ling as a cautionary example tied to perceived motion quality, while the production was described as pursuing “film-grade” real-time rendering to improve virtual being performance credibility.
Tourism: Jimo News (极目新闻) described an XR night-tourism deployment at Mulan Night City in Huangpi, Wuhan, where a “digital Mulan” was used as a destination-facing virtual being to accompany visitors during the holiday period, linking the character to local intangible heritage narratives and positioning the synthetic character as a centerpiece of a new immersive experience rather than a passive screen element. Harbin Ice and Snow World content dissemination included videos explicitly labeled as using Huiboxing digital humans, indicating the operational use of digital human presenters as the delivery layer for visitor advisories and venue-status messaging, with the virtual presenter functioning as a standardized front end for public-facing updates in a high-traffic winter tourism context.
Marketing: Baidu (百度) introduced an “all-chain” e-commerce solution in Macau centered on Huiboxing digital-human livestreaming, positioning the virtual being as the core interface for merchant acquisition and conversion workflows, and framing the deployment as a packaged solution intended to help local businesses reach customers more precisely while accelerating small and medium enterprise growth through digital-human commerce operations. Sina Tech (新浪科技) reported consumer-facing New Year greetings in Guangzhou that relied on algorithmic animation of old photos and digital-human performance to deliver personalized messages, treating the virtual being as the expressive layer that “performs” user-authored content rather than as a generic announcer, and explicitly linking the effect to an AI-enabled personalization workflow.
Enterprise: Huayi Zhiguang (华翼智光) launched the public beta of its brand digital human “Moliya” as a named virtual ambassador, with the deployment framed as a branded synthetic character product rather than a one-off media effect and accompanied by a roster of collaboration partners spanning animation, cultural IP, and heritage commercialization. The partner list included Hippo Animation (河马动画), Jibai Culture (江西即白文化传媒有限公司), Shangyi Media (上亿传媒股份), Luxiangyuan Group (露香园集团), Yunhui Technology (云麾科技), the Zhang Songmao Master Porcelain Panel Painting Art Museum Songmao Kiln (张松茂大师瓷板画艺术馆松茂窑), and Zhejiang Thangka Art Museum Treasure Pavilion (浙江唐卡艺术馆藏宝阁), indicating a strategy where the digital human is intended to travel across multiple content and cultural-product contexts, including intangible-heritage brand storytelling and IP-based promotional production. In Beijing, the 2026 Haidian New Year technology temple fair programming was described as including digital humans alongside large-model AI and immersive technology exhibits, placing virtual beings into a public science-and-culture event context where they function as demonstrators and experiential interfaces within a broader municipal showcase of AI applications.
Telecommunications And Devices: TPV Display Technology (冠捷显示科技) filed a patent application for a method enabling a virtual human image in voice calls, specifying a workflow where call audio is captured, encoded, compressed, transmitted over the network, and used to drive a cross-terminal virtual avatar under low-bandwidth conditions, which positions the virtual being as a live representation layer for communication rather than as prerecorded media. Asentech (亚盛科技) described an exhibition demonstration combining its local-deployment AI infrastructure prototype “Edge AI Array” with a virtual human demo running on that infrastructure, emphasizing on-device or edge-side deployment as the execution context for the synthetic character and presenting the virtual being as a visible application endpoint for showcasing the underlying computing stack.
February 19 News
Enterprise: Hunan Bank Co., Ltd. (湖南银行股份有限公司) issued a public tender for a “digital human system” construction project for 2025 with a reported budget figure of RMB 1.05 million, indicating an enterprise deployment context where a bank is procuring a packaged digital-human capability rather than a one-off media production. Migu Culture Technology Co., Ltd. (咪咕文化科技有限公司) and China Mobile Communications Group Co., Ltd. (中国移动通信集团有限公司) filed a patent application for a digital-human facial-expression driving method intended to improve real-time performance and expression naturalness, pointing to work on low-latency animation control and expression fidelity suitable for live-facing business scenarios such as service desks, hosts, or guided interactions. Tuguan (Tianjin) Digital Technology Co., Ltd. (图观（天津）数字科技有限公司) filed a patent application for a method and system for digital-human generation and interaction that explicitly targets realism, smoothness, and real-time responsiveness in interactive sessions, aligning with enterprise needs for stable, high-frequency conversational and presentational interactions in controlled environments. In a public-procurement governance context linked to multi-ministry guidance, the use of digital humans was also described as an assistive mechanism in bidding and tender evaluation, including “digital human defense” style presentations to help procuring entities compare shortlisted candidates, which frames digital humans as structured decision-support interfaces rather than entertainment characters.
Marketing: In Guangzhou’s high-volume e-commerce context described as a “Taobao village,” digital-human livestreaming was presented as part of an AI toolchain spanning product selection, script generation, and automated live presentation, positioning digital humans as scalable on-camera sales avatars used to raise throughput and reduce dependence on individual human streamers. Huawei Cloud (华为云) was referenced as part of a service ecosystem in which digital humans and “matrix systems” are integrated into a closed loop from technology to scenario to measurable results, indicating platform-led packaging of digital humans alongside multi-account content distribution and performance optimization. Guangzhou Yuntu Technology (广州云图科技) was described in the same orbit of optimization services, implying that digital humans are being embedded into broader marketing operations stacks where intent-tagging, content automation, and distribution tooling sit alongside synthetic presenters, with the digital human acting as the front-end persona for repeatable promotional content.
Tourism: Taishan Culture and Tourism Group (泰山文旅集团) launched an “AI companion tour” product described as an AI digital-human tour guide that provides real-time explanations and supports interactive use during peak holiday travel, framing the digital human as a persistent guide for on-site navigation, cultural interpretation, and visitor Q&A. In Chengdu, visitors at Du Fu Thatched Cottage were reported interacting with an AI digital human “Du Fu,” which situates a named historical-persona avatar inside a museum-like visitor flow as a conversational or interactive exhibit layer. In Shanxi’s cultural-tourism promotion framing, a digital human was positioned as a destination spokesperson for Qixian, emphasizing gesture-level performance and stylistic presentation designed to stand in for a human presenter in tourism marketing and local cultural storytelling. In Guangxi, an “AI smart exhibition hall” approach was presented as a set of solutions and project cases for building AI-enabled exhibition venues that include digital humans, implying deployments where a digital human functions as a venue receptionist, explainer, or guided-tour persona within curated physical spaces.
Entertainment: Multiple Spring Festival gala contexts described digital humans as on-stage synthetic performers and virtual hosts, including cases where virtual hosts appeared alongside real presenters, and where large ensembles of AI-generated digital humans were used for a group performance segment, indicating batch production pipelines for consistent character rigs, motion, and staging. Zhejiang Television’s virtual host “Gu Xiaoyu” was presented as part of a mixed-reality hosting format, where a virtual presenter anchors transitions and interacts within themed segments, reflecting broadcaster-led virtual-host branding rather than a generic avatar. Stagecraft descriptions also referenced real-time engine-driven interaction between live performers and AI-generated entities using Unreal Engine 5, which places digital humans and synthetic characters inside a real-time rendering and cue-driven production workflow rather than purely pre-rendered VFX. In the same entertainment environment, the virtual idol “Luo Tianyi” was referenced as a performance presence within a gala-style production context, reinforcing the continued use of named synthetic performers as recognizable stage acts within mainstream variety programming.
Cultural: In Jinan, cultural-heritage storytelling was described through a case where Longshan black pottery was personified as a digital human that performs rap-style narration, a format that converts museum or artifact information into character-driven performance and positions the avatar as a cultural interpreter rather than a neutral docent. In Shaanxi, Shaanxi Radio and Television Media Convergence Group (陕西广电融媒体集团) was associated with applied work combining AIGC and XR, and related cultural-program segments described ultra-realistic digital humans integrated into televised staging where virtual figures and real people intercut across different times and spaces, emphasizing compositing, spatial staging, and plausibly lifelike digital-human representation as part of cultural programming. In Xi’an’s immersive Tang-themed environment, named digital personas such as “Chang’an Guide Attendant” and “Chang’an Etiquette Attendant” were described providing consultation, explanations, interactive engagement, and companion-style services, framing the avatars as role-based synthetic characters that deliver themed guidance and visitor interaction inside a managed cultural venue setting.
Ethical Contexts: Social uses of “AI doubles” were described as being applied to real interpersonal situations such as annual-meeting performances, handling family questioning during holiday gatherings, and rehearsing dialogue for matchmaking contexts, treating the AI-generated persona as a stand-in that absorbs social load and enables controlled practice of self-presentation. Alongside these concrete scenarios, the same context explicitly raised risks of rights infringement and misuse, and highlighted the blurring of boundaries between real identity and synthetic representation, which are operational concerns whenever a digital human is used as a proxy for a specific person in semi-public communication. These concerns connect directly to deployment choices in marketing and entertainment where likeness, authorization, and downstream redistribution determine whether a digital persona is treated as a permitted character asset or a potentially misappropriated representation.
February 18 News
Entertainment: Keyiyun (客易云) was described integrating the Keling API (可灵API) to upgrade interactive digital humans used in video-generation workflows, positioning these humanlike synthetic characters as on-screen presenters and conversational front-ends for generated content across scenarios such as classroom-style “intelligent teaching assistants,” finance-oriented “virtual advisors,” and entertainment “digital idols.” Separately, large-scale televised gala production was described using XR stage techniques to support ultra-realistic digital humans presented as “traveling” across multiple scenes, and a related production case described “live-action + AIGC interaction” coupled with Unreal Engine 5 from Epic Games to enable real-time interaction between performers and AI-generated digital creatures, framing the virtual being as a co-performer embedded in the live show’s narrative and staging. (News reports link the “live-action + AIGC interaction” Unreal Engine 5 gala case to Jiangsu TV’s 2026 AI Litchi Spring Festival Gala, with Alibaba’s Qwen app described as the AI capability partner and Epic Games’ Unreal Engine 5 cited as the real-time engine.)
Cultural: Huairou Museum in Beijing was described upgrading its permanent historical exhibition while adding 3D virtual exhibition experiences that combine high-resolution images, audio-video content, and a digital human component that functions as an interpretive guide within the virtual tour context. In Chengdu, the “3791 Beidou Application Science and Technology Museum” (3791北斗应用科技馆) was described as featuring a digital-human dialogue area as a focal interactive installation, using named scientist personas for visitor conversation-style engagement inside the exhibition environment, aligning the digital-human interface with informal science education and on-site cultural participation.
Marketing: A regional “GEO optimization” case in Tianjin was described linking local search visibility improvements to a combined workflow that includes AI image generation and digital-human livestreaming, treating the virtual presenter as a sales-oriented interface that supports search-driven discovery and downstream conversion. A separate business case in Yinchuan was described using digital-human livestreaming for a home-furnishings manufacturer to drive online-to-offline traffic and incremental orders, presenting the virtual host as a scalable, repeatable sales channel embedded in short-video and livestream commerce operations rather than a one-off brand asset.
Enterprise: Yizhi Intelligence (一知智能) was described as providing AI digital-human services at enterprise scale, emphasizing multilingual delivery and accent or dialect simulation as a product capability used to better match the speech style of target audiences, with the company positioning repeated iteration and failure-driven refinement as part of building robust, customer-ready digital-human offerings. Beijing Yuedong Shuangcheng Technology (北京悦动双成科技有限公司) was reported as receiving a granted patent for an “AI animated digital human” apparatus and method oriented toward “face-to-face” communication with people, framing the digital human as an interactive agent designed to support direct conversational exchange rather than a purely pre-rendered avatar.
Platforms And Tooling: Soul App (Soul) was reported to have open-sourced a real-time digital-human generation model named SoulX-FlashHead, characterized as lightweight and intended to enable responsive, on-device or low-latency digital-human rendering in practical deployments. Xiaoduo Intelligent Technology (小哆智能科技) was reported filing a patent application focused on digital-human lip-sync generation using a “temporal decoupling” approach, describing a pipeline that includes face detection and cropping on original video sequences as part of synchronizing facial motion with speech, which positions the technology as a foundational module for digital-human video production and animation.
Social And Communication: Tencent (腾讯) was described rolling out a QQ feature called “QQ Brainstorm Show” (QQ脑洞秀) inside its Yuanbao (元宝) experience, enabling users to upload personal photos to generate Q-style virtual avatars and then produce short animated videos via templates, positioning the avatar as a consumer-facing creative instrument for everyday social posting. In multiple unrelated news-video contexts, Huiboxing Digital Human (慧播星数字人) was credited as the synthetic presenter used to narrate or deliver the video segment, reflecting a deployment pattern where a named digital-human presenter is used as a standardized production component for short-form news packaging and voiceover-style delivery.
Governance And Intellectual Property: Douyin (抖音) executives were cited discussing Seedance 2.0 in connection with digital “clones,” including constraints that the system does not currently support real-person face reference or established IP-character generation, and describing sustained investment in anti-infringement strategy as a major operational focus for this category of virtual-being content. In parallel, a creator-facing narrative framed digital-clone production as requiring real-person verification for generating a personal likeness, treating identity control and rights-management as built-in procedural requirements that shape how synthetic characters are created, moderated, and distributed on large-scale short-video platforms.
February 17 News
Cultural Heritage And Museums: Du Fu Thatched Cottage promoted a 2026 Spring Festival program that explicitly used a digital human persona of Du Fu (“数字人‘杜甫’”) as a visitor-facing interactive element alongside other “technology-empowered” experiences. In a separate museum-focused context, EasyAR (易现科技) was described as positioning its metaverse spatial computing stack EasyAR Mega as an enabling platform for AR guidance products that combine AR navigation with digital-human-style interaction, with examples framed as “digital human tour guide” offerings in museum guide deployments.
Entertainment And Virtual Idols: A Beijing Daily Client (北京日报客户端) New Year greeting video was released by YURI, presented as an AI virtual idol (“AI虚拟偶像”) with a claimed “digital identity,” built with generative AI by Hanqing Studio (汗青工作室) and operating under the AI.TALK brand, which was reported with concrete performance indicators including an audience scale described as over 1.1 million followers across platforms and a debut MV exceeding 12 million plays.
Tourism And On-Site Travel Companions: Taishan Cultural Tourism Group (泰山文旅集团) was reported as launching “Taishan Companion” (“泰山伴游”), positioned as an “AI companion tour” service implemented through an AI digital human to guide or accompany visitors as a new touring mode. In Shenzhen’s Gankeng Ancient Town (甘坑古镇), a holiday deployment was described as extending an earlier trial featuring an AI digital human “Li Bai” (“AI数字人李白”) into a broader “AI travel buddy” experience, where visitors book and use Rokid (Rokid) Glass 3 smart glasses as the front-end for the on-site companion interaction.
Public Transit And Commuter Services: Shenzhen Metro (深圳地铁) was described as running a cross-scenario interactive experience that included an online question-and-answer pathway through a NuwaAI mini-program, explicitly allowing riders to ask questions of a digital human interface as part of the commuter and holiday travel experience design.
Healthcare And Clinical Front Desk Triage: Seyo Technology (世优科技) was presented in a hospital deployment narrative as providing an AI digital human “doctor” used for 24/7 intelligent guidance functions such as directing patients, handling triage/diversion-style routing, and reshaping front-desk navigation flows in outpatient settings, with Nanyang Central Hospital (南阳市中心医院) named as the concrete on-site context where the digital human was described as operating in the outpatient hall.
Education And Campus Assistants: Seyo Technology (世优科技) also appeared in an education deployment framing that described an “AI digital human all-in-one” used as a full-scenario campus assistant, with the product name “Shiyou Bota” (“世优波塔”) given as the digital human system used for classroom-to-campus-history-museum style guided explanation and schoolwide assistance scenarios rather than a single-purpose kiosk.
Enterprise Assets And IP-Backed Finance: Agricultural Bank of China (中国农业银行) and Tianjin Bank (天津银行) were described in a data-asset audit/finance context as having issued loans where a “digital human” intellectual property certificate (“数字人’知识产权证书”) was approved as the pledged/collateralized asset in financing transactions, positioned as an example of data- and IP-like resources being treated as financeable assets in formal lending.
"Digital Human Intellectual Property Certificate" (数字人’知识产权证书) refers to an official-style IP registration certificate used in China to document and evidence rights or interests tied to a specific “digital human” asset, typically by registering the underlying data and/or digital content that constitutes the digital human (for example, its image/likeness materials, voice data, motion/action data, and related training or feature data) under emerging “data IP” or related registration mechanisms so that ownership, scope of use, and licensing can be evidenced and enforced; in practice it is being treated as an asset-like credential that can support commercialization and even financing, including reported cases in Tianjin where a “digital brain” digital-human certificate was pledged in a data-IP pledge loan transaction with banks, reflecting how such certificates are used to help turn digital-human-related data and IP into recognizable, transferrable, and financeable rights in business contexts.
Digital Human Production Tools In Media Workflows: Baidu Huiboxing (慧播星) was referenced as the digital human generation tool used to produce multiple short-form videos in unrelated news items, functioning as a workflow component that substitutes a synthesized human presenter for narration and on-camera delivery, indicating routine use of packaged “digital human presenter” tooling in mainstream content production.
Patents And Technical Development: Xiaoduo Intelligent Technology (小哆智能科技) was reported as filing a patent on a method and system for generating lip-synced speech-driven facial motion for digital humans, framed around a “temporal decoupling” approach to align speech and mouth-shape/face motion generation, with the work positioned as part of a broader digital human technical stack rather than a single branded character deployment.
February 16 News
Public Services: Fuzhou Public Security Bureau launched Fujian’s first police-service 3D digital human named Rong Xiao’an (榕小安), deployed inside the “Fuzhou Public Security” WeChat Official Account as an entry point called “Digital Police,” where residents can open an interactive page and communicate with the character in real time by voice or text; the described capability focus is semantic recognition and understanding for frontline public-service Q&A and convenience services delivered through a familiar mobile channel operated by Tencent (腾讯). In Xiong’an New Area, the Xiong’an Library placed a “smart digital human” greeter and recommender named Tu Xiao’an (图小安) on a large entrance display, positioned explicitly as an on-site guide that welcomes visitors and recommends books, illustrating a kiosk-style deployment of a digital human for public cultural facilities. (The technical support company behind Xiong'an Library—including the deployment of the smart digital-human librarian Tu Xiao’an (图小安)—is Xiong’an Chinese Online Digital Technology Co., Ltd. (雄安中文在线数字科技有限公司).)
Entertainment: Henan’s 2026 Spring Festival Gala removed human hosts and a physical stage and instead used AI virtual scenes to connect programs end-to-end, relying on AI animation transitions that became a visible point of audience debate about pacing and continuity; the concrete case is a full-show hosting-and-staging substitution in which a virtual environment replaces both the presenter role and parts of the stage language. In Guangxi’s 2026 Spring Festival opera program, a “digital human Liu Sanjie” (数字人刘三姐) appeared as a featured guest within a regional cultural showcase, using a well-known local cultural symbol as the basis for a synthetic character presented alongside location-based imagery such as the Yangshuo Yulong River segment. (Reporting attributes the build to the program’s own creative/production team within the CCTV Arts Program Center (中央广播电视总台文艺节目中心) using AIGC and virtual-space visual-effects workflows. It is one organizational unit within China Media Group, formally “中央广播电视总台文艺节目中心,” Arts and Entertainment Center of China Media Group.)
Media And Journalism: Beijing Daily’s client app and Jingbao Mobile Media (京报移动传媒) released a New Year greeting video featuring YURI, described as an AI virtual idol built with generative AI by Hanqing Studio (汗青工作室) under the AI.TALK label, framed as a “virtual person with a digital identity” delivering a formal holiday message; the reported specifics include an established fanbase and music-video distribution, positioning the virtual idol as a media-facing synthetic character used for festival greetings and brand-adjacent cultural content. Jinan staged an “all-AI-generated” gala hosted by a pair of digital hosts named Beichen (北辰) and Sixuan (思萱), presented as animated news-presenter-style characters crossing into entertainment hosting, which serves as a concrete example of digital humans replacing traditional host roles across a complete program rather than appearing as a single segment.
Tourism And Cultural Heritage: Shandong’s Tai’an upgraded “smart cultural tourism” services by putting an AI digital human tour guide online and adding a “smart companion travel” service oriented around Mount Tai tourism, explicitly tying the virtual guide role to a broader intelligent-travel product concept that raises innovation-versus-privacy tradeoffs as part of deployment considerations. (Mount Tai’s AI digital-human tour guide and “smart companion travel” service were developed by Tai’an Yuan Taishan Digital Technology Co., Ltd. (泰安市元泰山数字科技有限公司), a Taishan Culture and Tourism Group subsidiary.) Chengdu’s three-year national pilot plan for cinema and “smart cinema” (智慧影院) upgrades explicitly mentions using AI agents and virtual digital humans to improve service experiences, alongside “film + VR/AR” integration, describing a venue-service use case in which virtual beings and agentic systems become part of customer assistance and interaction inside cultural consumption spaces rather than being limited to screen content.
Enterprise And Finance: Scienjoy (思享无限) reported that its subsidiary Scienjoy Meta Technology (sjverse.com) launched an “AI Vista Live!” platform in Dubai on December 11, 2025, described as integrating multimodal AI with digital-human technology and targeting expansion in Middle East markets, making the digital human component a core interaction layer within a live-platform product strategy rather than a standalone demo. Xiaoying Technology (小赢科技) described “virtual digital human services” as part of technology-driven risk-control upgrades used to respond to broader credit-pressure conditions, positioning a virtual-being interface as an operational capability within financial services workflows rather than a purely marketing-facing avatar.
Telecom, Devices, And Consumer Assistants: Guangdong Mobile (广东移动) and ZTE (中兴通讯) jointly introduced the ZTE Starry 70 Ultra (星悦70 Ultra) as an “AI celebrity digital-human phone,” pairing built-in capabilities described as memory evolution, a digital clone, and one-utterance task fulfillment with an embedded AI spokesperson/digital-human persona named Yuan Yiqi (袁一琦), framing the device as a consumer endpoint for persistent assistant-like interaction through a branded synthetic character. Tencent (腾讯) added a consumer-facing avatar-creation feature called “QQ Brainstorm Show” (QQ脑洞秀) to the Yuanbao platform (元宝), where a user uploads a single photo to generate a chibi-style 3D avatar and then produces short AI videos via templates, including the option to invite friends’ avatars to co-star, providing a concrete pipeline from selfie to 3D personal avatar to multi-avatar short-form content creation within a social product surface.
Social Products And AI Clones: The social product Elys was described as decomposing social interaction into an “AI clone first” flow in which other users’ AI doubles read and respond to posts before humans step in later to confirm and deepen the conversation, making the AI companion layer an explicit mediator of timing, attention, and conversational load in a consumer social network design. (Chinese financial reporting describes Elys as an “AI-native application” released for testing by a company called Natural Selection (自然选择) that was strategically invested in by Kingnet Network (恺英网络), and the same report links Natural Selection’s prior release of a 3D AI companion app (EVE) to the same product line.) Changzhuo AI (畅卓AI) was presented as a short-video “multi-pose clone” system that emphasizes face and pose reconstruction across viewpoints, describing a creator toolchain aimed at replacing or supplementing live performance with a controllable digital double for rapid short-video production.
Open Models And Production Technology: Soul App (Soul) released an open-source real-time digital-human generation model named SoulX-FlashHead, with performance framed around single-GPU (RTX 4090) feasibility for real-time head-generation workloads, providing a concrete example of a consumer social platform externalizing a digital-human generation capability as a reusable model artifact rather than keeping it exclusively proprietary. Huawei Cloud CodeArts (华为云码道 CodeArts) integrated a code-oriented intelligent agent with the open-source model GLM-5 from Zhipu AI (智谱AI), and this same news cluster explicitly positioned “agents” and open models as adjacent enablers for digital-human workflows, linking practical developer tooling for multimodal/agentic capability with the broader ecosystem that also includes real-time digital-human generation models.
February 15 News
Government & Public Services: In Gansu, a data-element industrial park project was described as planning to use artificial intelligence and digital humans to build a “digital culture-and-tourism matrix,” positioning digital-human experiences as part of a province-wide cultural tourism offering and linking this to the wider build-out of large-scale computing capacity in places such as Qingyang where utilization demand is a practical concern. In Hubei, the Ezhou municipal government described introducing an AI digital human called “E Xiaobang” (鄂小帮) into a “one-stop” enterprise service reform pilot, with continuous learning over government-service knowledge and pro-business policy to provide real-time voice interaction and fast localized responses for consultation-style requests.
Culture & Heritage: In Shanghai, BesTV (百视通) and Shanghai Museum (上海博物馆) were described as bringing an online “Ancient Chinese Sculpture Gallery” experience to a broadcast/streaming context, with a production approach that includes high-definition 3D generation of flat cultural relics, dynamic extension of scroll paintings, life-oriented reconstruction of wearable jewelry relics, and the use of digital-human techniques to support “clothing and adornment” style restoration so that historical scenes and cultural stories can be presented with a more embodied, explainable layer rather than only static images. In Chongqing, a library “smart experience area” was described as using a digital-human persona to re-present historical figures from the War of Resistance period and narrate related stories, alongside an adjacent interactive “one-click costume change” experience framed as a user-facing application of AI for historical ambience and engagement.
Entertainment: Live Performance, Festivals, and Events: In Beijing, the Haidian New Year technology temple fair was previewed as a multi-point, multi-exhibit event explicitly including three digital-human points, multiple virtual-reality/augmented-reality enhancement points, large-model interactive experiences, and digital interactive games, treating digital humans as one category within a broader interactive exhibition program rather than a single centerpiece. In Hunan, a Spring Festival gala program description referenced an AR digital human appearing in performance staging with a singer, treating the digital human as an on-stage co-present element enabled by AR compositing. In Guangxi, a program description for a nationally broadcast Spring Festival opera gala referenced a “digital human Liu Sanjie” (数字人刘三姐) appearing as a featured guest, using a regionally iconic character as a synthetic performer identity for broadcast entertainment.
Marketing & Commerce: In Beijing, “AI New Year greeting cards” (AI拜年咔) were discussed as a viral consumer sharing format in which AI-driven personification and templated production are used to generate greeting content designed for reposting within social feeds, and it was framed as a step toward broader consumer adoption of digital-human-style creation tools. In Harbin, a case story described an individual using AI digital-human technology as a stand-in to keep a livestream running while the original host traveled home early for the holiday, presenting digital humans as a practical substitution mechanism in creator-led commercial attention workflows rather than only as a novelty effect. In Guangzhou, a marketing services case described a solution combining “tap-to-trigger viral splitting,” an AI digital human, and unattended livestreaming for a local chain hotpot business, with the digital human used as the front-facing on-stream presence integrated into a growth and conversion loop rather than as a standalone avatar experiment.
Enterprise & Developer Platforms: Soul App (Soul) was reported as open-sourcing a real-time digital-human generation model called SoulX-FlashHead, positioned around lower-cost deployment of higher-quality real-time digital humans, including performance claims framed around single-GPU inference and streaming responsiveness, and presented as an enabling component for developers or small teams building real-time digital-human applications. Tencent Cloud (腾讯云) was described as providing an “agent development platform” used across a large share of leading cultural media institutions, with described scenario coverage including cultural-tourism guidance, intelligent content creation, cultural Q&A, and digital-human interaction, indicating a platformized approach where digital-human interfaces sit alongside other NLP and large-model capabilities as part of an integrated media and cultural-production toolchain. Baidu Search (百度搜索) was described as releasing an “open” real-time interactive digital-human agent, explicitly framing the digital human as an interactive “agent” layer rather than a purely rendered figure, and positioning it as an industry-facing capability for real-time, human-present interaction.
AI Companions & Synthetic Characters: A report framed “AI pets” as a growing form of “digital creature” used by young consumers in China for emotional comfort, characterized by always-available interaction and low-burden disclosure dynamics compared with living pets, which situates AI companions as a consumer application class adjacent to but distinct from humanlike digital humans, with the “pet” form used to structure attachment, routine interaction, and private conversation. (Within China, two prominent, clearly virtual (software-first) company examples that map to the AI/virtual-pet niche are Zhejiang Jinke Tom Cat Culture Industry (浙江金科汤姆猫文化产业股份有限公司), whose portfolio includes large-scale virtual pet-raising mobile games in the Talking Tom & Friends franchise (a virtual-pet category leader by distribution and downloads, even when not framed as “AI pets” in the newer emotional-companion sense). and LumiAni (灵秘科技), which is explicitly positioned as building a hyper-realistic virtual pet application platform using NeRF-based neural rendering and AI generative models, with its current emphasis described as a mobile virtual-pet app.)
Identity, Digital Twins, and “Digital Self” Systems: Zhongbei Telecom (中贝通信) and Wuhan Dagong Intelligent Technology (武汉大公智能科技) were reported as jointly launching a digital-avatar debating platform called “AI72 Debate” (AI72辩) in Wuhan, alongside a core “ByteMe” digital-avatar system, described as aiming to give professionals a “personal AI assistant” style capability through an individualized digital self that can argue and respond, and described as relying on an internal technique characterized as “micro-alignment” to move from generic tool behavior toward user-specific cognition and stance representation. Tianya Community (天涯社区) was described as packaging “build an AI digital-human clone” as part of a paid “founding member” revival plan tied to account data recovery and long-term access benefits, treating the digital-human clone as a premium identity extension product linked to legacy user data rather than a general-purpose creative tool.
Legal, Governance, and Consumer Protection: Beijing Internet Court (北京互联网法院) was referenced as establishing a digital-asset collegiate panel and as having mediated multiple disputes involving digital-human infringement, placing digital humans within a judicially recognized digital-asset and rights-protection context and signaling that courts are developing procedural capacity for disputes around identity likeness, model outputs, and related IP claims. A consumer-protection advisory in Hubei treated AI virtual-human livestreaming as a disclosure and compliance issue by warning that virtual-human livestreams should be clearly labeled as AI-generated, framing synthetic presenters as a risk vector for misleading marketing if identity and generation status are obscured.
Technologies and IP/Patents: TPV Electronics (冠捷电子科技) was reported as obtaining a patent for a display approach “matching” AI digital humans to enable more efficient and user-friendly human–machine interaction, placing digital humans directly into hardware-interface optimization rather than only content creation. XinshiYing Technology (新时颖科技) was reported as applying for a patent on apparel digital-human model generation and parameter optimization with an explicit goal of maintaining multi-platform adaptation effects, treating the digital human as a retail/creative asset that must render consistently across platform pipelines. Guangyu Zhilian (Changchun) Technology (光屿智联（长春）科技) was reported as applying for a patent on an encrypted interaction method and system for digital humans designed to resist common cracking attacks, framing digital-human interaction as a security surface with explicit cryptographic threat modeling. Inspur Smart Technology (浪潮智慧科技) was reported as applying for a patent on a cross-modal identity-authentication method for virtual-human interactive control, emphasizing identity verification across modalities as a prerequisite for controlling or authorizing interactions with a virtual-human interface in multi-sensor environments.
February 14 News
Healthcare: Chongqing Bishan District’s grassroots “digital-intelligence empowerment” project (重庆市璧山区数智赋能基层项目) explicitly includes “digital human” function scenarios as part of a CNY 35 million build intended to upgrade local primary healthcare service capability, and the AI-avatar model is also illustrated by the China-based medical app Ant Afu (蚂蚁阿福), where doctors provide consult-like services through clinician-linked AI doubles that users can query for triage-style guidance and medication direction, framed as a way to route scarce expert capacity toward remote or underserved areas.
Culture And Museums: Chongqing Library (重庆图书馆) deployed an on-site “digital librarian” (数字馆员) inside an immersive Chongqing War of Resistance memory experience, combining a digital human interface with interactive “one-click outfit change” and AI-animated wartime comics that “move” within the exhibit space, while Chengdu Du Fu Thatched Cottage Museum (成都杜甫草堂博物馆) used a digital human version of Du Fu (数字人“杜甫”) to converse with visitors during its 2026 Spring Festival programming and Jinan’s “Tianxia Quancheng AI Spring Festival Gala” (2026天下泉城AI春晚) staged a digital human “Li Qingzhao” (数字人“李清照”) as part of an AI-driven cultural performance lineup that re-presents historic figures through synthetic character embodiments.
Tourism And Transit Services: iFLYTEK (科大讯飞) and Shaanxi Cultural Industry Investment Holdings Group (陕西文化产业投资控股集团) are described as advancing culture-and-tourism deployment through localized large-model landing plus virtual digital human development, and Shenzhen Metro (深圳地铁) is described as running a Spring Festival “tech show” in which a digital human performed Cantonese greetings and also functioned as an “intelligent travel advisor,” positioned as a jointly built public-facing service experience with the Longgang District Culture, Radio, Television, Tourism and Sports Bureau (龙岗区文化广电旅游体育局).
Enterprise And Government Operations: The Ministry of Industry and Information Technology’s “Industrial Internet Platform High-Quality Development Action Plan (2026–2028)” (推动工业互联网平台高质量发展行动方案（2026—2028年）) explicitly names “intelligent inspection digital humans” (智慧巡检数字人) and process-automation assistants as targeted capability directions for industrial platform ecosystems, and in a separate enterprise-product direction JD Technology (京东科技) describes “JD AI Pay” (京东AI付) as a JoyAI-model-enabled payment mode intended to run across AI assistants and smart terminals and already realized in “universal digital human” scenarios, presenting digital humans as a software carrier for transaction guidance and execution rather than as a purely presentational avatar.
Marketing And Consumer Education: Bank of Communications Sichuan Branch (交通银行四川省分行) described a consumer-protection engagement pattern built around “digital human + quiz gamification + prize draw,” reporting large-scale participation as a measurable outcome and positioning the digital human as the interactive front-end that delivers compliant financial literacy content in a street-life themed public setting, while Huamedia Holdings (华媒控股) referenced a subsidiary practice case involving Hangzhou Net’s “Hangzhou Wenbo Gold Medal Recommender No. 1” digital human (杭州网“杭州文博金牌推荐官1号”) as an example of a media-linked branded digital human spokesperson/recommender concept.
Education And Training Pipelines: Jiangxi Xinhua Computer Institute (江西新华电脑学院) is described as offering an AIGC film-and-animation design curriculum that treats digital human modeling as a core applied module alongside AI script generation, AI animation rendering, and virtual scene design, framing digital humans as a standard production skill inside an integrated content pipeline rather than as a standalone novelty.
Synthetic Models, Open Source, And Enabling Tech: Soul App (Soul App) is reported as releasing an open-source real-time digital human generation model branded as SoulX-FlashHead (SoulX-FlashHead) and also reported under the closely named SoulX-FlashTalk (SoulX-FlashTalk), with performance and scale details such as a 14B-parameter profile and sub-second latency claims, positioning it as a technical substrate for live digital human synthesis, Ant Group (蚂蚁集团) is described as open-sourcing Ming-Flash-Omni 2.0 as unified audio-generation infrastructure explicitly tied to downstream domains including digital humans and content creation, Kepuyun (开普云) is described as applying speech-recognition capability into digital-human interaction platforms, and China Mobile (中国移动) appears in a patent-framed case in which a user’s digital-asset data is used to generate a unique digital human image and then produce a “memory video,” describing an end-to-end synthetic character artifact assembled from personal data inputs.
E-Commerce Fashion Assets: Fujian Xinshi Ying Technology Co., Ltd. (福建新时颖科技有限公司) and Fujian Xinshi Ying Brand Operation Co., Ltd. (福建省新时颖品牌运营有限公司) are identified as applicants on a patent describing clothing digital human model generation plus parameter optimization intended to safeguard cross-platform adaptation effects, treating the digital human as a reusable commercial modeling asset whose controllable parameters must remain stable across multiple distribution environments.
Identity, Personal Clones, And Data-Backed Personas: ByteDance (字节跳动) is linked to consumer apps Doubao (豆包) and Jimeng (即梦) that reportedly require audio/video real-person verification before users can generate a personal digital human “clone” (数字人分身) and then use that verified synthetic self to produce AI video, and Tianya Forum (天涯论坛) is described as packaging “build an AI digital human double” (构建AI数字人分身) within a paid founding-membership concept alongside long-range data recovery and digital identity badge benefits, presenting digital humans as a premium identity product anchored to historical account data.
Ethical Contexts And Consumer Harm Patterns: A concrete harm narrative is supplied by the case of “AI domineering CEO” accounts (AI霸总AI数字人账号) described as “hunting” older women through persuasive synthetic-persona content, illustrating how digital humans and synthetic characters can be operationalized for manipulation and targeted emotional exploitation, and a compliance-oriented counterpoint is supplied by the described rule that AI virtual-human live commerce should be clearly labeled as AI-generated, positioning disclosure as the minimum mitigation layer when synthetic presenters are used in transactional persuasion.
February 13 News
Healthcare: Chongqing Bishan District Medical and Health Affairs Center (重庆市璧山区医疗卫生事务中心) planned a “digital-intelligence empowerment at the grassroots” program with a stated budget of 35 million RMB that explicitly included digital-human capabilities as part of its functional scenario set, positioning “digital humans” as an applied component within "information and intelligent construction related to primary healthcare" (基层医疗相关的信息化与智能化建设) rather than a stand-alone showcase. Ant A Fu (蚂蚁阿福) operated an app-based health service flow that combined question answering with downstream digital healthcare steps and added “famous-doctor AI avatars” (名医AI分身) led by six academicians together with more than 1,000 doctors, enabling remote users—explicitly including rural township users—to consult through mobile access to these expert-derived AI personae for health education and consultation.
Education And Law: Guangxi Minzu University School of Law (广西民族大学法学院) provided a 3D digital-human interface that students could “ask” about ASEAN legal questions, framing the digital human as an interactive legal-knowledge access layer rather than a passive display; in parallel, Nanning Railway Transport Intermediate Court’s Nanning International Commercial Tribunal (南宁铁路运输中级法院南宁国际商事法庭) was referenced as a setting where disputes could proceed on appeal, situating “digital-human interface Q&A” within a broader legal-services and cross-border legal-literacy context even when the adjudicative venue itself was not described as virtualized.
Commerce And Marketing: Suqian Dasu Agriculture Development (宿迁大树农业发展有限公司) described a livestream-commerce workflow that relied on short, intensive training cycles and pointed toward exploration of “digital humans” as a next step in upgrading livestream formats, alongside the broader encouragement for enterprises to apply big data and AI to innovate livestreaming operations at the regional level. A Harbin livestream seller case described substituting an AI digital human to “stand in” for a livestream host to allow the human operator to travel home for the Spring Festival, showing a concrete use case of avatarized hosting as continuity staffing for routine commerce broadcasting. JD.com (京东) launched “JD AI Pay” (京东AI付) and tied it to JoyAI large-model capabilities, describing applicability to AI assistants and intelligent terminals and stating that it already supported ordering food inside a “universal digital-human assistant” JoyAI app experience via a “universal doctor” (万能博士) interaction path, treating the digital human as the conversational front end for payments and transaction execution.
Culture And Tourism: China Media Group (中央广播电视总台) presented the digital human “Liu Sanjie” (刘三姐) as a featured guest element for its 2026 Spring Festival Opera Gala programming, using the AI digital figure to perform or introduce content associated with the well-known song “The Folk Song Is Like Spring River Water” (山歌好比春江水), and thereby using a synthetic character to bridge televised stagecraft, regional cultural symbolism, and program narration. China Unicom (中国联通) was associated with an immersive “Millennia Yingge” (千年英歌) experience hall concept that used a digital-human image and interaction mechanics where visitors could drive a digital Yingge performance formation to move with drum rhythms, explicitly shifting the visitor role from passive viewing to participatory control of a virtualized cultural performance. Ningxia highway service areas (宁夏高速公路服务区) were described as adding an AI digital-human interactive guidance system as part of a broader “convenience experience” upgrade, using a virtual guide to support wayfinding and service inquiry in a travel-and-rest setting rather than a museum or stage environment.
Public Services And Infrastructure: Zhangjiakou municipal government service facilities (张家口市政务服务中心) were described as deploying a “digital-human police officer” (数字人警察) to answer public consultations at the public-security counter, framing the virtual persona as a front-desk triage and Q&A layer for routine citizen services. (They explicitly state that the system is an in-house build by the Zhangjiakou Municipal Public Security Bureau, describing it as the “Zhang Jingguan” (张警官) large-model–based AI agent and presenting the digital human in the service-hall as one of its application scenarios.)
Museums And Heritage Presentation: A Shanghai museum-going scenario described the Shanghai Children’s Museum (上海儿童博物馆) as using multiple digital technologies—explicitly including interaction with “metaverse digital humans”—to let visitors navigate immersive environments and engage virtual personae as part of the exhibition experience. BesTV (百视通) and Shanghai Museum (上海博物馆) were associated with a “cloud” presentation model for the museum’s Chinese Ancient Sculpture Gallery that combined high-definition 3D generation of artifacts, dynamic expansion of scroll paintings, and lifestyle reconstruction of wearable artifacts, and also described using digital-human techniques to realize “clothing and adornment” presentation functions, treating a digital person as the carrier for historically grounded apparel visualization rather than only presenting objects in isolation. A Beijing cultural project titled “Digital Storytelling of Intangible Heritage in Beijing’s Grand Canal” (《“数说”京城大运河里的非遗》) was described as using AI digital-human technology to enliven and present content tied to the Temple of Heaven’s Divine Music Office traditions (天坛神乐署中和韶乐), positioning the virtual figure as a medium for heritage interpretation and staged reenactment.
Consumer Social, Avatars, And Short-Video Creation: Tencent (腾讯) introduced a QQ feature called “QQ Brainhole Show” (QQ脑洞秀) that let users upload a single photo to generate a personal 3D avatar for short-video creation and then invite friends’ avatars to co-star, framing avatars as a lightweight co-creation primitive that embeds AI generation into everyday social video workflows rather than requiring professional pipelines.
Enterprise And Industrial Digital-Human Systems: Hunan Bank (湖南银行) was referenced as running a “digital human system construction” tender in the context of broader technology-organization restructuring, implying a bank-internal deployment where a digital human functions as a front-end service or interaction layer within financial operations. Hangzhou Grand Canal Digital-Intelligence City (杭州大运河数智城) highlighted local-company achievements in AI-enabled manufacturing and noted that Xiangxin Technology (相芯科技) received recognition for an “AI digital-human agent” (AI数字人智能体), framing the digital human explicitly as an agentic system rather than only a visual host, and placing it within an industrial and regional innovation-evaluation context.
Platforms, Models, And Tooling For Real-Time Digital Humans: Soul AI Team (Soul AI团队) was cited as having open-sourced a podcast voice-synthesis model called SoulX-Podcast and a real-time digital-human generation model called SoulX-FlashTalk, with SoulX-FlashTalk described with 14B parameters, 0.87-second low latency, and 32 fps high-frame-rate output, positioning the stack as enabling responsive, real-time synthetic presenters or characters across speech and singing-related scenarios rather than only offline rendering. Quwan Network (趣丸网络) applied for a patent describing a dialogue-generation method designed to maintain long-range consistency of role characteristics using a preset persona-trait lexicon (人设特征词库), treating “digital humans” as persistent characters whose conversational identity must remain stable across extended interactions. Beijing Ditingshijie (北京谛听视界) described an “emotional-level” holographic digital-human AI interaction terminal called the DT-BOX, positioned as a home-oriented device intended to bring a holographic digital persona into domestic settings through an interactive terminal form factor.
Risks, Abuse, And Governance Contexts: A case report described “AI CEO/romance” style digital-human accounts (AI霸总AI数字人账号) that targeted middle-aged and elderly women by cultivating attachment and then inducing victims to leave contact details, after which the operators sold health products and online courses, with one cited case involving an elderly woman paying 2,000 RMB, concretely framing synthetic-character deployment as a fraud and manipulation vector rather than a benign companion. Separately, an anti-fraud publicity package described a “digital-human anti-fraud song” as one component of a themed set of public messaging assets, indicating that digital humans are also being used as attention-optimized carriers for public-safety communication in response to such risks.
Media Production And Digital Presenters: Baidu Huiboxing (慧播星) was explicitly credited as the digital-human presenter tool used in multiple short-form video items, treating the “digital human” as a production utility for narration or on-camera delivery in news-like clips rather than as the subject of the story, which concretely illustrates how virtual presenters are being inserted into routine media workflows as a replacement or supplement for human hosts.
February 12 News
Healthcare: Shanghai-based researchers described a “digital prophet” approach for precision medicine that uses AI to simulate extracellular vesicles via a construct referred to as AIVEVs (“artificial intelligence virtual extracellular vesicles”), positioning it as a four-in-one pipeline for prediction, simulation, diagnosis, and design that can connect digital biology workflows with clinical translation, and noting its visibility through publication in the journal Bioactive Materials alongside the claim that these AI-modeled vesicle systems could function as an enabling bridge from computational modeling to medical application.
Public Services And Safety: Chongqing Anti-Fraud Center (重庆市反诈骗中心) and Duxiaoman (度小满) launched a Spring Festival anti-fraud package that centers a digital propaganda image ambassador presented as an AI-native virtual idol, Yuri Youli (尤栗), with the concrete media output being an MV for an anti-fraud song titled Didu Guardian (嘀嘟守护) released across the internet and complemented by additional AI-driven anti-fraud content such as short-form dramas; in Guangzhou, the municipal “Guangzhineng” (广智能) platform was reported to include a policy-interpretation digital human, Sui Xiaozhi (穗小智), positioned to explain and interpret local “AI + industry” policy and related measures; at the national administrative-process layer, the National Development and Reform Commission (国家发展和改革委员会) and co-signing ministries and agencies referenced introducing “digital human Q&A/defense” mechanisms in bidding and tendering to support evaluation of shortlisted bidders, record the full decision process, and improve traceability of award decisions.
Education And Tourism: In Nanjing, a primary-school initiative framed as “AI accompanies me to tour Jinling” presented a localised AI education and cultural-transmission setup built around an IP character rendered as a digital human, Mingdangshen Xiaobao (明党参小宝), which was described as providing voice-based interaction while guiding learners through Nanjing-specific cultural points of interest; Juyou Technology (巨有科技) was reported in the context of study-tour “AI companion” systems that combine education and tourism by using AR plus interactive digital humans to convert static exhibits into dynamic, conversational knowledge carriers; Guangxi Tourism Development Group (广西旅游发展集团) was described as operating a Liu Sanjie digital human (刘三姐数字人) that is integrated into the One-Click Tour Guangxi (一键游广西) platform for tourism guidance and service delivery; additional tourism-facing experimentation was referenced through a NuwaAI (NuwaAI) digital human embodiment of the poet Li Bai being deployed as an AI-enabled attraction in a cultural tourism setting; Yuanmeng Space Cultural Tourism Technology (Chengdu) Co., Ltd. (元梦空间文旅科技（成都）有限公司) was reported as providing a “one-stop, full-region cultural tourism metaverse construction solution,” including an AI digital human generation-and-interaction system it had already released in 2022, with that system referenced as having been considered within a Universiade-related metaverse project track.
Entertainment And Media: Zhewen Interactive (浙文互联) was described as producing short-form drama content built around multiple virtual digital humans, including Tianyu (天妤) as the focal point of a short-drama series titled Qianbi Xunzong (千壁寻踪) that was reported to have surpassed 400 million views within a year of launch, alongside additional named virtual digital humans Jun Ru Jin (君若锦) and Lan LAN (兰LAN) as part of its broader slate; in televised variety programming, an AI virtual host named Xiao Mage (小马哥) was reported to have appeared in the 2026 Chengdu–Chongqing children’s Spring Festival gala recording, interacting live with on-stage hosts via banter and improvised dance as a concrete example of real-time virtual-host participation; separately, a major-stage audiovisual performance context was described where virtual imagery/avatars were used as part of stagecraft to switch character presentation and support transitions between virtual and real staging in a Yue opera excerpt.
Marketing And Commerce: A news video segment about celebrity-assisted produce sales explicitly noted the use of Baidu Huiboxing (慧播星) digital human technology as the on-screen presenter mechanism for delivering the content, illustrating a concrete pattern where a digital human “front end” is used to package and narrate commerce-related messaging in a broadcast/social-video format rather than relying on a human presenter for the full delivery.
Technology, Platforms, And Tooling: Soul App (Soul) announced the open-sourcing of a real-time digital human generation model named SoulX-Flash Talk, specifying performance targets including sub-second latency (0.87s), 32fps output, and long-duration stability at a 14B scale, and separately described additional open releases in the same family such as SoulX-Podcast (podcast voice synthesis) while also reporting a joint release with Geely Automobile Research Institute AI Center (吉利汽车研究院人工智能中心（AIC）) of SoulX-Singer as an industrially oriented zero-shot singing voice synthesis model; Jidong Pangu (吉动盘古) disclosed a patent application for a WebGL-based system for synchronising a digital human mouth model and customising scenes, describing a workflow that combines a digital human 3D model, lip-sync/viseme animation, and scene elements to enable user-facing visual interaction with a digital human inside web-rendered environments.
Ethical Contexts: ByteDance (字节跳动) was reported to have adjusted access controls around Seedance 2.0 in response to trust and safety concerns linked to highly realistic “digital double” capabilities, including limiting real-person face reference pathways and applying platform-level interception of requests that involve real-person facial inputs on specific web and platform surfaces while requiring biometric verification on mobile—implemented by recording a user’s own image and voice to create a dedicated personal digital double—positioning identity verification, face-input restriction, and capability limiting as operational controls for managing realism-driven misuse risk in video-generation workflows.
February 11 News
Standards And Governance: The Ministry of Industry and Information Technology of the People’s Republic of China and the Ministry of Science and Technology (中华人民共和国科学技术部) of the People’s Republic of China (alongside the Ministry of Housing and Urban-Rural Development of the People’s Republic of China, the State Administration for Market Regulation, and the China National Intellectual Property Administration) issued a 2025 edition guide for building a standard system for the technology services industry, setting a target of drafting 40+ new national and industry standards by 2027, with reporting that SenseTime (商汤科技) is leading the drafting of a national standard for virtual digital humans, situating “digital humans/virtual humans” as an explicit standardization object within a government-led services and innovation framework.
Entertainment And Broadcast: Entertainment reporting describes a Spring Festival Gala scenario in which a human host introduces a “digital partner” onstage, with emphasis on high-detail facial skin texture and hair dynamics, framing the virtual being as a co-host; related coverage around the CCTV Spring Festival Gala (央视春节联欢晚会) also includes discussion of audience reactions to a human host paired with a digital human presenter, treating the synthetic character as a visible on-camera role rather than a behind-the-scenes tool.
Cultural And Heritage: Broadcast and local reporting around the 2026 CCTV Spring Festival Opera Gala (央视春节戏曲晚会) states that a digital human version of “Liu Sanjie” appears as a special guest, presented not as a static reconstruction but as an AI-driven performing and interactive digital character tied to a well-known cultural IP, with the program described as using virtual-space visual effects and AIGC methods; separately, local cultural-policy commentary in Nanning proposes building a smart-museum platform and developing Guangxi-culture-themed AI digital human guides, explicitly positioning digital humans as museum interpreters and (in the same proposal set) extending guidance needs to include ASEAN-related visitor contexts.
Consumer Devices And Assistants: Retail-device coverage describes a large-screen “AI phone” model, WIKO X70, being marketed with a built-in assistant app called Dongdong App (东东) characterized as a “universal digital human assistant” that combines intelligent Q&A, health inquiry/triage-style questioning, takeaway ordering and shopping support, and creative generation functions; the app is described as belonging to JD.com (京东) and is framed as a digital-human-style helper designed for older users’ daily smartphone use, with the “digital human” element positioned as the interaction front-end for multiple utility functions.
Enterprise Content Creation And Platforms: Platform and industry items describe multiple China-based deployments where digital humans function as content presenters and interactive fronts: 263 Network Communications (二六三) is described as having launched AI digital humans named “Yun Xiaoduo” and “Chang Xiaoping” based on its “263 metaverse” capability and RTN (Real-Time Network), with stated landed scenarios including virtual hosting and 3D showrooms; Mingmang (Beijing) Intelligent Technology (明芒（北京）智能科技有限公司) announced the global launch of an AI video website called aistudioslab (aistudioslab.com), described as centering “global digital humans” as core capabilities for short-drama and broader content creation; Taobao Vision (淘宝Vision) is reported to use a self-developed “3D Gaussian real-person digital human” technique for real-time rendering and spatial interaction in a “future fitting room” scenario; and ByteDance (字节跳动) is referenced in policy/usage reporting about Seedance 2.0 (seedance2.ai) imposing restrictions on real-person face uploads across named products (including requirements that users record their own likeness and voice for real-person verification before producing a “digital avatar/digital double”), indicating an identity-verification gate as part of consumer-facing digital-double creation workflows.
Healthcare: Consumer-health reporting describes an “AI medical-visit assistant” use case in which “digital doubles” are used operationally, including a stated 2025 initiative in Shanghai’s Huangpu District health administration (上海市黄浦区卫生健康委员会) to create simulated digital doubles for 200+ real family doctors that replicate each doctor’s voice characteristics, with coverage indicating the intent is to support services for a population described at roughly 400,000 residents, implying a clinician-facing synthetic persona layer for patient interaction and access to primary-care services.
Education: Education reporting on Beijing University of Posts and Telecommunications describes a “future learning building” with multi-modal smart classrooms and labs where AI agents and digital humans are embedded into classroom teaching, framing digital humans as part of instructional delivery and in-class interaction infrastructure rather than as standalone media characters.
Public Safety And Legal: Public-safety reporting states that the Chongqing Anti-Fraud Center and Du Xiaoman (度小满) jointly launched a Spring Festival anti-fraud campaign package that includes an AI-native virtual idol named Yuri Youli as an anti-fraud “digital image ambassador,” plus the release of what is described as a first domestic “digital-human anti-fraud song” MV titled “Didu Guard” (“嘀嘟守护”) targeted at younger audiences, alongside additional anti-fraud content such as an AI anti-fraud short drama and materials focused on deepfake prevention; in parallel, legal commentary on virtual idols and VTuber-type creators discusses infringement-liability questions in the context of “doxxing/opening boxes” targeting a virtual persona, treating the virtual being as a rights-bearing identity in disputes involving exposure of private information and reputational harm.
February 10 News
Entertainment: Bilibili (哔哩哔哩) surfaced an AI virtual host named Qiuzhi (秋芝) as a prominent on-screen presenter in its Spring Festival Gala programming, with audience discussion centering on the virtual host’s performance, perceived stage presence relative to human celebrities, and the practical limits of scripted delivery highlighted by an incident in which Qiuzhi reportedly misread “quantum computing” as a near-homophone before continuing after a brief pause; in parallel, a virtual streamer identified as DL.Du (DL.毒) circulated a heavily re-edited dubbing performance of a Spongebob “pickle” scene, with attention focused on the synthetic/performative voice-over style and live comment reactions as the primary engagement mechanism; within state-broadcast entertainment formats, China Media Group (中央广播电视总台) programming also featured “digital human” co-hosting concepts discussed by viewers in relation to a human host pairing with a highly detailed virtual counterpart on a Spring Festival stage, treating the digital performer as an on-air partner rather than a background visual effect.
Marketing: A cross-border e-commerce “one-person company” workflow was described around Baidu Keevx, presented as an application capable of learning high-performing short-video patterns, reproducing a similar style, and generating “digital human” marketing videos tailored to overseas aesthetic preferences, combined with a one-click multilingual pipeline framed as 178-language translation to localize the resulting promotional videos for different markets, positioning the synthetic presenter as a scalable front-end for small-team marketing output rather than as a single fixed character tied to one brand. (Keevx is an AI “digital human” (avatar) video-creation and video-localization product operated by Baidu (百度), presented through a Baidu-branded service endpoint and footer/legal identifiers that explicitly attribute the service to Baidu.)
Tourism: Taishan Cultural Tourism Group (泰山文旅集团) was described as deploying an AI digital-human tour-guide capability and an accompanying “smart companion tour” service built around Mount Tai’s natural scenery and historical-cultural materials, with concrete functions including historical-figure-style narration for the site, scene-based interaction, and itinerary planning modules, presented alongside a broader “smart tourism” stack that explicitly integrates AR/VR and IoT elements to support mixed physical–digital visitor experiences anchored by the digital-human guide as the interface layer. (The technical implementer most directly tied to the “AI digital-human tour-guide / smart companion tour (AI escort)” capability you described is Taian Yuan Taishan Digital Technology Co., Ltd. (泰安市元泰山数字科技有限公司), a subsidiary of Taishan Cultural Tourism Group that the Mount Tai Scenic Area Administrative Committee credits with independently developing a “scenic-area AI companion-tour solution based on a cultural-tourism large model,” explicitly characterized by “historical-figure companion touring” and already deployed across dozens of points of interest at Mount Tai.)
Cultural: A named synthetic character, “Digital Human Liu Sanjie” (数字人刘三姐), was introduced as a special-guest presence for the 2026 Spring Festival Opera Gala ecosystem associated with Guangxi International Communication Center (广西国际传播中心), with the use case framed as applying AI digital-human techniques to re-present a canonical regional cultural figure within a televised performance context; separately, Huamei Holdings (华媒控股) attributed an award-recognized deployment to its wholly owned subsidiary Hangzhou Net (杭州网), identifying a specific digital-human implementation called “Hangzhou Wenbo Gold Medal Recommender No.1” (杭州文博金牌推荐官1号) and noting that it received a first-prize designation in an AI-application case selection run by China City News Websites Alliance (中国城市新闻网站联盟), positioning the digital character as a public-facing recommender/presenter for city cultural-content promotion.
Enterprise: China Telecom Qinghai (中国电信青海) was described as launching a WeCom-based (enterprise WeChat) digital customer-service capability to deliver “palm-sized” service access, with an explicit plan to introduce digital-human technology to create a regionally styled virtual customer-service persona as the front-end interface; in a related tooling-and-supply context, Laihua Technology (来画科技) was characterized as an animation and digital-human intelligent generation platform emphasizing controllable AI video output, positioning production infrastructure for synthetic presenters/characters as a reusable capability rather than a single one-off character.
Finance: China Life Yunnan Branch (中国人寿云南省分公司) described using “digital human” content formats within a Spring Festival travel-season financial-education campaign, including distribution of risk-education and consumer-protection videos that explicitly included a “digital human” segment alongside other explainer formats; ICBC Jilin (中国工商银行吉林) was described at a high level as deepening digital-human technology application within “digital finance” work to reduce repetitive labor and shift staff attention toward higher-value tasks, treating the digital human as an operational interface for internal and customer-facing workflows rather than as entertainment media.
Ethical Contexts: A rights-and-liability dispute was described around a virtual digital human referred to as “Moubao” (某宝), with the legal reasoning presented as treating reputational-rights protection as belonging to civil subjects while construing the synthetic character as an AI product without civil legal status and therefore not itself bearing a reputational right, even where contested statements were directed at the character and circulated in public channels; the case framing, as presented, positioned virtual beings as objects of dispute in defamation-like narratives while anchoring enforceable claims in the rights of natural or legal persons rather than in the virtual character as an independent rights-holder. (The case write-up anonymises the character as “某宝” and does not formally name the developer, but the described project most plausibly points to Kugou Music (酷狗音乐), as the company behind the virtual idol.)
Kugou Music is a long-running Chinese online music and audio entertainment platform that operates as one of the major apps within Tencent Music Entertainment’s portfolio, combining streaming discovery, social music features, and music-centric live interaction through products such as Kugou Music and Kugou Live; in the digital-human context, Kugou has positioned itself not only as a distributor of music but also as a technical and operational base for synthetic performers, using in-house voice synthesis and AI-singing infrastructure branded as the Lingyin Engine (凌音引擎) to support “AI singers” and virtual idol releases, and it is publicly associated with virtual-idol operations linked to the SING ecosystem, including the virtual member Shanbao (扇宝) that is commonly described as being produced under Kugou-affiliated cultural/studio branding (for example, Qigu Culture/Studio, 齐鼓文化/齐鼓工作室) and active across platforms such as Bilibili and Kugou’s own live-streaming channels, illustrating Kugou’s role as both a platform and a developer/operator layer for music-oriented digital humans.
February 09 News
Healthcare: Gushengtang (固生堂) launched an “AI clinic” model described as “AI avatar + offline young doctors collaborative diagnosis and treatment,” where patients can connect across regions to named-doctor AI avatars for traditional Chinese medicine access and training acceleration for younger clinicians, with the example that a patient in Guangzhou can connect to AI avatars of doctors based in Beijing and Shanghai so that care in one city can reach nationally distributed expertise; Naodong Aurora (脑动极光) signed a cooperation agreement with Kiang Wu Hospital (澳门镜湖医院) to implement cognitive digital-therapy and brain-health digital AI technical services in Macau as a concrete institutional deployment of health-focused digital/AI human technologies.
Entertainment: The 2026 China Network Audio-Visual Gala (2026中国网络视听盛典) was reported as using AIGC, holographic projection, and virtual humans as core production techniques to stage a program positioned around “culture and technology” integration, presenting a concrete broadcast case where virtual-beings and synthetic-media pipelines are treated as on-air creative infrastructure rather than experimental add-ons.
Marketing: 3D Communications (三维通信) was reported as pursuing an “AI marketing + commercial space” dual-engine strategy and, in the marketing lane, referenced collaboration involving Alibaba (阿里巴巴) e-commerce virtual-human livestreaming and AIGC content-material workflows framed around data-compliance barriers and authoritative platform endorsement, with a concrete operational datapoint tied to Taobao (淘宝) “Flash Sale Day” where peak daily orders were reported at 120 million and the virtual-human/AIGC cooperation was positioned as a mechanism showing how compliance and content governance are treated as part of the virtual-being commercialization stack in high-volume retail campaigns.
Enterprise: Sitiqi Information Technology (思特奇信息技术股份有限公司) was reported as winning a China Telecom (中国电信) procurement for a Beijing 2025 program described as “AI agent applications and digital visualization construction R&D (digital human),” presented as a concrete operator-side deployment in which a digital-human front end is explicitly coupled to agentic large-model applications and visualization engineering as part of a carrier’s production technology roadmap; Yizhi Intelligence (一知智能) was presented in a Zhejiang innovation-ecosystem report as having a named “digital-human product” associated with its productization efforts, serving as a concrete example of a startup-style digital-human offering being advanced within a local incubation and policy-support environment rather than only within major platform firms.
Creative Tools: Wondershare Technology (万兴科技) was reported as upgrading the desktop version of its video-creation product suite to deepen AI-assisted production and extend coverage across multiple creative domains explicitly including digital humans, alongside adjacent capabilities such as AI video generation, video conversion, AI comic content, and related tooling, positioning “digital human” outputs as part of an integrated creator workflow rather than a standalone novelty feature.
Infrastructure & Identity: Nokia (诺基亚) was reported as having a China patent publication for a “5G-trust-based virtual avatar authentication” method, described as verifying virtual-avatar use within application functions based on trust anchored to the user device, providing a concrete identity-and-authentication case where “virtual avatar” is treated as an authenticatable security object in telecom-grade architecture rather than merely a front-end presentation layer.
February 08 News
Government Services: Guangzhou Basic Geographic Information Center (广州市基础地理信息中心) was recognized at a Guangzhou-hosted Guangdong digital government forum in connection with four newly awarded industry group standards that explicitly include an “AI digital human application specification,” signaling formalization of how digital humans are designed, deployed, and evaluated in public-sector contexts. Zhaoqing High-tech Zone (肇庆高新区) was cited as a long-running provincial digital-government innovation case and as a participating drafting unit for a “government service hall AI digital human application specification,” linking digital humans directly to on-site service-hall workflows where they function as front-desk interfaces for consultation and guidance. In a separate municipal planning context, Liangxi District’s government service flagship store deployed a government-service digital human named Xiaoxi (小溪) as an online-facing service persona within a broader push to apply large models in public administration, and Jinan’s big-data authority demonstrated a policy-release scenario in which a senior official’s digital human “twin” delivered a bundled policy announcement to support “AI super-individual” entrepreneurship, illustrating the use of avatarized officials for standardized, repeatable public communication.
(Zhaoqing High-tech Zone is an industrial development zone in Zhaoqing, Guangdong, that has positioned “digital government” as a practical delivery system for business-facing and citizen-facing services, and its digital-human work sits inside that broader service-modernisation effort rather than as a standalone media or entertainment initiative: local authorities have reported that the zone’s government office, administrative service center, and a local data-services company were named participating drafters for the group standard “政务服务大厅人工智能数字人应用规范” (T/DGAG 042—2025), a provincial-level standard released through Guangdong Digital Government Association and led by Guangzhou Government Service Center, which enumerates common service-hall digital-human scenarios such as reception, policy explanation, consultation and guidance, assisted form-filling, assisted acceptance and approval, and online services. In parallel, the zone has publicised the rollout of an AI assistant integrated into its “government and business service platform” mini-program, describing the deployment of DeepSeek capabilities to improve “enjoy without application” policy matching and to streamline review/approval steps in policy-disbursement workflows, framing the technology as a way to reduce manual burden and speed up policy delivery for firms and residents.)
(“Application Specifications of Artificial Intelligence Digital Humans in Government Service Halls” (T/DGAG 042—2025) is a Guangdong group standard that formalizes how AI “digital humans” are used inside government service halls as a public-service interface, and it does this by turning digital-human deployment into a requirements-and-assessment problem rather than a branding or vendor novelty problem. When it specifies what systems should cover in practice, it is effectively defining a baseline service catalogue and operational behaviours for a hall digital human—what kinds of interactions it must support (for example, reception-style guidance, consultation, policy explanation, and assisted handling support), what the outputs must look like in service terms (such as directing users to the right counter, clarifying eligibility rules, or guiding form completion), and what integration expectations exist with the hall’s online and offline workflows. When it specifies how deployment quality should be evaluated, it is establishing criteria and an evaluation method that can be applied across deployments—so a site can test whether the digital human reliably provides correct guidance, performs consistently in real hall conditions, meets usability and service-effectiveness expectations, and can be operated and maintained to an agreed standard. The practical consequence is that governments can write clearer tender requirements, vendors can be compared against the same measurable expectations, and supervisors can audit performance over time, reducing the risk that each service hall ends up with an incompatible, unmeasurable, or purely cosmetic “digital human” that cannot be justified or scaled.)
Consumer Protection And Regulation: Guangdong Provincial Market Regulation Administration (广东省市场监督管理局) and Guangzhou Municipal Market Regulation Bureau (广州市市场监督管理局) were described as advancing a “trusted consumption livestream room” group standard that explicitly brings AI-generated content—such as digital human livestream hosts—into the compliance perimeter, with requirements that livestreams using these techniques meet governance expectations oriented toward protecting consumers’ legal rights and strengthening safeguards for minors. In parallel, entertainment-industry dispute reporting framed “digital doubles” as an enforcement and identity-verification problem in livestream commerce, describing a scenario where an actor discovered dozens of impersonating AI personas simultaneously selling products and faced the practical burden of proving authentic identity in order to pursue rights protection, positioning synthetic-character misuse as a concrete driver for clearer rules on performer-related AI use and platform accountability.
(Guangdong Provincial Market Regulation Administration is the provincial market regulator that, in the context of digital humans, is presented as both a consumer-protection authority and a compliance driver for AI-enabled livestream commerce and public communication: it warns against deceptive marketing practices that use AI-generated virtual presenters and simulated “experts” to fabricate credibility, it is associated with ongoing public-service communication that uses digital-human presenters for continuous outreach such as food-safety education, it appears in compliance-oriented efforts that treat “digital human” livestream selling as a risk area alongside broader livestream e-commerce issues, and it is linked to emerging expectations that platforms operationalize rules such as negative lists for livestream marketing while implementing specific “AI digital human livestream compliance” measures, including experimentation with interactive “AI legal-education digital humans” for public law-popularization.)
(Guangzhou Municipal Market Regulation Bureau is Guangzhou’s local authority for supervising advertising, consumer protection, and market order, and in the digital-human context it functions as the municipal layer that turns national livestream e-commerce supervision into practical compliance expectations and enforcement, treating AI-generated “digital human” livestream presenters as accountable marketing actors rather than a special case; this means digital-human livestreaming is expected to fit into full-chain controls such as identity/qualification verification, clear responsibility allocation among platforms and operators, truthful and non-misleading promotion standards, record-keeping and auditability, and effective complaint handling, with the regulator positioned to issue local guidance, conduct inspections, and pursue enforcement where digital-human use contributes to deceptive claims or consumer harm. )
Enterprise Training And Customer Service: Mofa Technology (魔珐科技) was presented through concrete productized deployments that tie virtual beings to routine enterprise operations: a virtual-livestream AIGC platform enabling around-the-clock live commerce or branded livestreaming with AI-driven and real-time voice-driven control; a 3D virtual-human service line oriented toward scalable, service-oriented deployments; and a workflow that turns slide decks into micro-lesson videos so training content can be produced and refreshed faster, with additional emphasis on AI training-video and digital-human training-video use cases. iFLYTEK (科大讯飞) was described as offering a virtual-anchor video tool designed for low-friction production—selecting a virtual persona, entering text, and generating a presenter-style video—mapping virtual beings to standardized internal communications, product explainers, and other templated narration tasks where speed and repeatability are central.
Healthcare: Gushengtang (固生堂) was described as launching an “AI clinic” model that operationalizes “AI avatar + offline young doctor” collaboration, enabling patients in Guangzhou to access the AI avatars of expert traditional Chinese medicine physicians from Beijing, Shanghai, and other cities while local clinicians handle in-person execution, thereby treating the virtual being as a front-line expert interface for cross-regional resource sharing and clinical capacity extension. (Gushengtang built and deployed its AI clinic and expert “AI avatar” system in-house by integrating it into its clinical HIS workflow, using Alibaba Cloud (阿里云) computing resources and Huawei (华为) Ascend AI chips, and formalizing technical cooperation with Huawei.) A separate example framed an AI “double” used to resolve routine, basic health questions at scale so that hospitals and physicians can concentrate on higher-acuity treatment, with Mayi Afu (蚂蚁阿福) cited as a named system in that role and associated with very large reported usage since its mid-2025 launch, illustrating the AI-companion pattern as a triage-and-guidance layer rather than a replacement for clinical care.
Culture And Tourism: A Kunming cultural-tourism project built around the Xu Xiake IP implemented an “AI digital human + real-scene interaction + intangible cultural heritage experience” pattern to activate historical and cultural resources through immersive participation, treating the digital human as an in-situ guide and interaction layer inside a physical attraction rather than a purely online asset. (Yunnan Jiangwu Culture Technology Co., Ltd. (云南讲武文化科技有限公司) delivered the Kunming Xu Xiake–themed culture-and-tourism experience “霞客游滇·大观寻章,” combining an AI digital-human “digital Xu Xiake” with on-site real-scene interactive elements and intangible cultural heritage activities to guide visitors through an immersive physical attraction.) In Yunnan’s Spring Festival travel context, Bank of China Yunnan Branch (中国银行云南省分行) was described as participating in a public-education activity that used a financial consumer-protection digital human, positioning the synthetic character as a standardized explainer for safety and rights education in high-footfall public settings.
Media And Entertainment: Zhejiang TV (浙江卫视) was described as deploying a virtual host named Gu Xiaoyu (谷小雨) for a Yue Opera-themed Spring Festival program lineup, illustrating the virtual-presenter use case in broadcast entertainment where continuity, branding, and format consistency are key. A national network audio-visual gala was described as integrating AIGC, holographic projection, and virtual humans to stage multi-screen spectacle, reflecting a production pipeline where synthetic characters are combined with spatial display technologies to extend performance grammar beyond conventional filming and stagecraft. In news production, a Henan Daily-affiliated visual all-media center was described as fielding a digital-human reporter team and continuing to generate additional digital reporters across beats including culture and tourism, demonstrating synthetic characters as reusable reporting and presentation assets rather than one-off special effects.
Technology And R&D: Zhejiang Shikongzhizi Big Data Co., Ltd. (浙江时空智子大数据有限公司) was linked to a disclosed patent application for an interactive digital-human multimodal presentation method that combines large-model RAG with prerendering-match logic, explicitly targeting faster, higher-quality multimodal interaction by aligning retrieval-augmented generation outputs with suitable prerendered assets for display. Soul App (Soul) open-sourced a real-time digital-human generation model, SoulX-FlashTalk, described as a 14B-parameter system achieving sub-second end-to-end latency, positioning low-latency synthesis as a practical enabler for live, conversational digital-human experiences rather than offline rendering. Baidu (百度) was described as announcing what it characterized as the first dual digital-human interactive livestream room alongside a “highly fused” multimodal digital-human approach, framing the technical direction as multi-agent or multi-character co-presence in live formats. Lianzhitong (Beijing) Technology Co., Ltd. (立安智通（北京）科技有限公司) was linked to a granted patent for a real-time digital-human streaming method using multi-process decoupling and dual-mode adaptive pushing, focusing on robustness and adaptivity in live delivery. Xiaoduo Intelligent Technology (Beijing) Co., Ltd. (小哆智能科技（北京）有限公司) was linked to a granted patent for a multimodal emotion-recognition-based digital-human interaction system, placing affect sensing as a core interaction primitive. Moore Threads Intelligent Technology (Beijing) Co., Ltd. (摩尔线程智能科技（北京）股份有限公司) was linked to a patent application targeting portrait-generation model improvements for finer digital-human facial detail in rendering and video synthesis. Renxin Network Technology Co., Ltd. (广州市人心网络科技有限公司) was linked to a patent that combines virtual-face selection model training with a privacy-preserving online psychological-consultation method and system, explicitly tying digital-human face selection to confidentiality protection in counseling workflows. Great Wall Motor (长城汽车) was linked to a granted invention patent for “virtual avatar display” methods and devices in vehicles, treating the avatar as an in-car interface element. Hanfan Digital Technology Co., Ltd. (天津汉梵数字科技有限公司) was described as building a digital-human system integrated with Huawei Cloud (华为云) ecosystem capabilities and connected to a WIT large model, positioning “large model + digital human” integration as a packaged capability for enterprise deployment. SenseTime (商汤科技) was referenced as leading the drafting of a national standard for a customer-service virtual digital human, tying virtual beings directly to standardized service quality and interoperability expectations. Unilumin Group (洲明科技) was described as entering a strategic cooperation with Mofa Technology (魔珐科技) oriented toward “AI digital human multimodal interaction,” framing hardware-display capability and digital-human interaction stacks as a combined pathway for next-stage deployments.
February 07 News
Government And Regulation: China’s industrial and cyberspace regulators were referenced in several items that frame virtual beings as regulated digital-service infrastructure rather than novelty media, including multi-agency guidance on improving inbound digital services for overseas visitors that sits alongside ongoing digital-identity and content-governance work affecting digital humans used in public-facing services and commerce, explicit treatment of live-streaming e-commerce where AI-generated presenters and “digital human anchors” are brought into the governance perimeter for generated content, and a Ministry of Industry and Information Technology (工业和信息化部) public consultation process for a compulsory national-standard project on metaverse classification and identification that includes requirements for “digital human identity identification,” implying a push toward standardized identity markers for digital humans used across platforms and scenarios.
Cultural And Tourism: China Youth Travel Service (中青旅) described a concrete tourism deployment at the Gubei Water Town scenic area (古北水镇) in which an in-house virtual digital human supports routine visitor interactions such as daily inquiries and product recommendations, while Juyou Technology (巨有科技) was cited for a cultural-heritage style venue installation in which a “Xiaoyou” digital human at the Tibet Film Culture Corridor (西藏电影文化长廊) relies on a customized knowledge base to narrate and explain a multi-decade film history in an interactive way, shifting the experience from one-way presentation to an on-demand explanatory guide; a separate venue case in Shenzhen described an on-site digital human assisting ticketing and entry, where Huawei’s HarmonyOS (鸿蒙) near-field communication and “atomic service” capabilities enable a phone-near-gate trigger that opens a passage credential for frictionless entry, coupled with AR real-scene navigation that guides visitors precisely through the site. (CYTS Tours Holding Co., Ltd. (中青旅) is a China-based travel and cultural-tourism group whose most concrete “digital human” deployment, as publicly described by the company, is at Gubei Water Town Scenic Area, where it has built an in-house virtual digital human by leveraging iFLYTEK Co., Ltd. AI capabilities to serve as a front-line, visitor-facing interface for routine inquiries, product recommendations, and voice interaction, with delivery designed to reach visitors via both large-screen and small-screen touchpoints on site.)
Public Services And Civic Communication: Multiple municipal and public-administration uses appeared as named case studies, including the launch of “Xiaochan” (小禅), presented as the first government-service digital human for Foshan’s Chancheng District (禅城区) designed to support interactive service delivery as part of a broader “smart government” model, and a “digital avatar” of Zhang Xi (张熙), introduced as the head of Jinan’s Big Data Bureau, used to deliver an on-stage release at the Jinan “digital-intelligence ecology” OPC community launch, framing the digital human as an official proxy for public communication and announcement delivery rather than only a customer-service interface.
Media And Language Interaction: A provincial broadcasting group showcase in Fujian highlighted a Minnan-dialect interactive digital human named “Yang Xiaohe” (杨小和) that conducts intelligent dialogue in Minnan (闽南语) and answers questions related to local audiovisual content and regional culture, positioning dialect-capable speech interaction and localized knowledge as the differentiator, while a legal-protection outreach case in Hubei described an AI digital human “Youyou” (佑佑) used as a digital carrier for youth-related legal education, supporting real-time interaction via voice and text as part of public-facing rule-of-law communication.
Commerce And Marketing: A large-scale commercialization claim was made for Zhejiang Wenhu Interconnect (浙文互联), which attributed substantial 2025 performance to its “Paizhi” digital human (派智数字人) as a revenue-generating digital marketing and digital-culture capability, and another consumer-facing personalization example appeared in Ant Group’s “Afu” app (蚂蚁阿福), where an “elder mode” upgrade adds the ability to customize the Afu persona’s appearance with selectable digital human styles and indicates ongoing expansion of avatar options; these cases emphasize digital humans as repeatable marketing assets and brand touchpoints, distinct from one-off promotional stunts.
Enterprise And Operations: Digital humans were also described as operational roles inside organizations and service workflows, including Agricultural Bank of China’s Jiangsu branch being referenced in connection with a centralized-authorization “digital human” review outsourcing procurement context, and a tax-administration case describing a “tax digital human” module inside a “big-data foundation and intelligent judgment” platform used by Tianjin tax authorities, presented as a functional component that supports structured administrative work rather than a purely conversational front end.
Healthcare And Education: Weizhong (唯众), Wuhan Weizhong Zhichuang Technology Co., Ltd. (武汉唯众智创科技有限公司), was referenced in a Changsha visit note that described a mental-health direction using a digital human paired with an “intelligent venting device,” where the digital human is said to run on an LLM-style large model plus a specialized psychological knowledge base to support psychological-service interactions, while Shandong Digihuman Technology (山东数字人科技股份有限公司) was named as the winning supplier for a Zunyi Medical University teaching-platform expansion project in Guizhou that upgrades anatomy and functional-laboratory facilities for a provincial basic-medicine experimental teaching demonstration center, presenting “digital human” technology as part of formal medical-education infrastructure procurement rather than a consumer app.
Technology, Platforms, And IP: Several items focused on enabling technologies and intellectual property for real-time, scalable digital humans, including Lianzhitong (Beijing) Technology (立安智通（北京）科技有限公司) receiving a granted patent for a real-time digital human streaming method described as “multi-process decoupling” with “dual-state adaptive pushing,” and Xiaoduo Intelligent Technology (Beijing) (小哆智能科技（北京）有限公司) receiving a granted patent for a digital human interaction system based on multimodal emotion recognition, indicating continued engineering emphasis on stable real-time delivery pipelines and affect-sensitive interaction; on the application side, Banmu Liangren Bioengineering Research (Ningxia) (半亩良人生物工程研究（宁夏）有限公司) applied for a patent describing a large-model-based digital-human live-stream interactive system for personalized agricultural product recommendation, tying foundation-model interaction to sales-oriented recommendation in a specific vertical.
Risks And Misuse: Two concrete risk patterns were described as expanding alongside adoption, including live-streaming e-commerce scenarios where AI digital humans can be used to impersonate public figures for sales fraud and are therefore being pulled into stronger oversight discussions, and employment and training scams that use “AI digital human job” narratives as a “recruit-then-train” inducement script, where the digital-human job label functions as the lure rather than the actual work content, indicating that virtual-being hype is being operationalized both for consumer deception and for labor-market fraud schemes.
February 06 News
Healthcare: Several China-based cases position digital humans and AI “splits” as front-door clinical interfaces and context-aware companions. In the report about Ant Group (蚂蚁集团), the “Afu” digital-human interface is described as gaining an elder-focused mode with large-type UI and voice calling, alongside newly opened customization of the Afu digital-human avatar so the on-screen persona can be tailored while remaining the interaction surface for care-related services. The Paper’s reporting on AI clone (“AI 分身”) describes a concrete rural use case in which an individual opened an “AI split” inside the Afu experience so herders can consult about health questions and medication guidance locally, framing the “split” as a consultative proxy that reduces travel for minor issues. Separately, a healthcare navigation case is described at Xi’an City Wall, where a “Tang-dynasty girl” digital human is paired with health monitoring: when a visitor’s heart rate is detected as abnormal, the system triggers an AR guide into a simplified mode and recommends nearby rest spots, combining cultural guidance with lightweight risk-aware assistance. In hospital operations, an AI digital-human named “Xiaoya Xi” is described as an intelligent triage/self-service kiosk put fully into use at Xiangya Jiangxi Hospital, illustrating the “digital human at intake” pattern as a scalable interface for routing and guidance.
Entertainment: Broadcast-stage and performance contexts emphasize digital humans as on-air presenters and as production assets integrated into major events. Coverage of the 2026 Yue Opera Spring Festival Gala describes a virtual host, “AI Gu Xiaoyu,” used to carry the program’s narrative device of time-travel presentation and to support the staging of opera skills within a “digital stage” concept, with the virtual host positioned as the connective performer rather than a mere graphic insert. Reporting about China Media Group / China Central Television (中央广播电视总台) indicates that its “CCTV Listen Media Large Model 2.0” was applied to Spring Festival Gala content production, alongside image-generation and “ultra-realistic digital human” techniques, explicitly treating hyper-real digital humans as part of a modern AIGC-enabled production pipeline rather than a novelty cameo.
Marketing: Multiple items frame virtual humans and digital humans as brand-facing growth instruments, with emphasis on measurable engagement and “new content form factors.” A China Press Industry item on GEO service providers describes the combined use of virtual-human technology and GEO content strategy to build an AI digital spokesperson for a consumer brand, explicitly claiming increased interaction and consultation volume on AI platforms while positioning the spokesperson as a youth/tech-image asset. Summit coverage by Yiou (亿欧) about “AI-driven brand growth” characterizes digital humans as a now-maturing marketing application that moved beyond early “stiff” visuals into deeper brand enablement, using the digital-human layer as a reusable growth interface rather than a one-off campaign render.
Tourism And Cultural Experiences: Several concrete China-based deployments treat digital humans as always-on explainers and trip assistants embedded in scenic areas and cultural programming. China Youth Travel Service (中青旅) states that its Gubei Water Town (古北水镇) scenic area built an in-house virtual digital human that handles daily Q&A, product recommendations, and voice interaction, describing a pragmatic service-avatar that sits between visitor intent and purchasable offerings. In Chengdu’s Spring Festival programming, “digital-human explanations” are described as part of public-facing cultural experiences that blend food, folk customs, and arts with mediated interpretation, positioning the digital human as a scalable docent. In Guizhou cultural tourism coverage, a provincial cultural-tourism digital human named “Huang Xiaoxi” is described as being integrated into a tech-enabled tourism experience, with the “digital guide” concept tied to personalization via user profiling and virtual-space construction to deliver “thousand people, thousand guides” style narration rather than a single fixed script.
Enterprise And Workplace: A set of cases describe digital humans and avatars as labor interfaces for internal operations, training, and high-throughput screening. Hunan-focused industry coverage describes “digital humans” conducting 24/7 interviewing, where entering a role requirement can generate test questions, presenting the digital human as a standardized, always-available front-end for recruitment workflows. A Jiangsu-focused item about a media-group strategy event describes stage use of “AI splits” and a “digital-human group,” including a named host-like digital human “Qinghe” and a medical-expert persona “Yan Xiaoyi,” portraying digital humans as institutional spokes-avatars and expert-character shells used for corporate communication and domain storytelling. In Zhejiang, an item about a “3D digital-human hall” describes rapid “one-second” replication to create a personal digital self that can converse and dance and is framed explicitly as something that can be passed down within a family context, placing the digital double in the category of identity continuity and household memory assets rather than only workplace productivity.
Platforms And Open Source: One of the most technically specific items is the open-sourcing of a real-time digital-human generation model by Soul App (Soul), described as SoulX-FlashTalk. The reports characterize it as a real-time, end-to-end system with sub-second latency (0.87s) and 32fps output, and they frame the release as pushing the sector toward commercialization by enabling “real-time online interaction” scenarios rather than offline rendered avatars; the key case implication is that a platform-native model release can standardize low-latency digital-human generation for interactive products, including real-time conversational presence and continuous output constraints.
Regulation And Governance: Multiple items explicitly connect virtual anchors and AI digital humans to compliance obligations, treating “synthetic presenters” as regulated actors in commerce and media rather than neutral graphics. Chongqing and Chengdu-focused reporting on live-streaming e-commerce governance describes new measures that bring AI digital humans and virtual hosts into the regulatory scope and require prominent labeling, presenting the policy logic as consumer transparency around generated personas used in selling contexts. Separate national-level and provincial references to “digital human broadcasters” being brought under supervision similarly frame the digital-human presenter as a governance object—defined by its ability to simulate a real host—so enforcement can target undisclosed generation, impersonation risk, and misleading performance in commercial or public-information settings.
February 05 News
Education: In Jiangsu, classroom teaching was reported to incorporate a historical-character digital human case in which the digital persona “Wang Rong” speaks in a childlike voice yet delivers mature lines and debates students role-playing as “Zhuer,” using real-time dialogue to make classical texts “speak” and to stage in-class argumentation as a learning activity; in Jinan, a school-focused “AI psychology companion” scenario was described in which students access article and video resources and interact with a customizable digital human called “Xiaoxing,” including avatar dressing and voice changes, positioning the avatar as a youth-facing interface for psychoeducation content and self-help engagement rather than a static content portal.
Cultural Tourism: In Chengdu, Spring Festival cultural programming was described as using digital-human explanations and guided experiences across major venues including Sichuan Museum, Chengdu Museum, Wuhou Shrine Museum, and Du Fu Thatched Cottage Museum, framing digital humans as on-site narrative agents for exhibits and holiday activities; a separate cultural-heritage exhibition format was described in which visitors conduct “naked-eye 3D” conversations with hyper-realistic 3D digital humans representing poets such as Li Bai and Du Fu; in Lhasa, Shiyou Technology (世优科技) was reported to have deployed an interactive AI digital human called “Xiao You” in a film-culture corridor setting, using a large-screen/holographic-style installation for exhibition interaction and visitor guidance.
Government And Public Services: Multiple public-service deployments were described as shifting digital humans from display to service roles such as consultation, guidance, processing assistance, and evaluation, including the “Guiren Zhiban” AI assistant example and the “Ai Shandong” Jinan service-hall capability example as representative “single-entry” service patterns; in Shanghai, Shanghai Hongqiao Railway Station was reported to provide 24-hour Q&A and indoor navigation through an AI digital human named “Xin Xiaolu”; in Xiong'an New Area, a library-lobby case described a smart digital human “Tu Xiao’an” used for visitor-facing interaction; in Taizhou, a grassroots people’s congress case described a digital human “representative” used to support localized duty performance and resident-facing engagement; in Xiangmihu, an international service “micro-window” case described an AI digital human providing multilingual greetings and front-desk-style assistance as part of place-based internationalization services; in Nanning, a China–ASEAN Expo setting described the public conversing with a digital human “Jiang Xiaomei,” and a policy suggestion was described to build an ASEAN multilingual AI consumer-service platform led by a regional science-and-technology authority in coordination with the municipal government.
Media And Broadcasting: In local media, Sanming Sanyuan Integrated Media (三明三元融媒) was described as putting an AI anchor “on duty” to deliver information via digital-human broadcast, emphasizing rapid delivery of on-scene updates through synthetic presentation; at an industry event, Fujian Provincial Radio and Television Group (福建省广播影视集团) was described as showcasing an interactive Minnan-dialect digital human “Yang Xiaohe” that conducts intelligent dialogue in Minnan, using regional language capability as the differentiator for on-site demonstration and audience interaction; at the national gala level, China Media Group (中央广播电视总台) was described as using hyper-realistic digital humans within an 8K, domestically sourced production workflow and applying “CCTV Listen Media Model 2.0” (央视听媒体大模型2.0) to content production, highlighting AI-assisted image generation and digital-human production as efficiency and quality levers for large-scale live entertainment.
Commerce And Marketing: In residential community programming, Midea Real Estate (美的置业) was described as introducing two AI virtual hosts—“Da Xu” and “Dandan”—for a multi-city owner gala, using scripted yet interactive hosting and “warm” transitions as the functional replacement for conventional emcees; in enterprise storytelling and branding, Sinopec (中国石化) was described as using digital humans alongside AI documentary production and AI composition in an internal creative “storytelling” event, treating synthetic presenters and AI-generated media as tools to modernize corporate narrative and stage performance; in product-promotion systems, Aiyu New Vision (Shandong) Digital Technology (艾渔新视界（山东）数字科技有限公司) was described as securing a patented “product promotion method and system” based on digital humans, positioning the avatar as the primary interactive marketing surface for presenting products and conducting promotion workflows.
Health And Wellbeing: A tourism-linked wellness case described a scenic-site digital human persona—framed as a “Tang-dynasty girl” character—combined with health-monitoring functions so that when a tourist’s heart rate is detected as abnormal the system can trigger an intervention, treating the avatar as both a narrative guide and a biometric-aware front end for safety-oriented service; in student mental-health support, the previously noted “AI psychology companion” scenario described the avatar “Xiaoxing” as the user-facing persona for engaging learning resources and self-guided support interactions.
Core Technologies, Platforms, And IP: Soul App (Soul) was described as open-sourcing a real-time digital human generation model called “SoulX-FlashTalk,” with reported sub-second latency (0.87s) and 32fps long-form video generation positioning it for real-time avatar conversation and live-like video interactions; Baidu (百度) was described as filing a patent for an agent-based interaction method intended to enable virtual avatars to interact intelligently with users in context-appropriate scenarios, explicitly spanning use cases such as metaverse-style interaction, smart customer service, and digital-human livestreaming; Dongxin Shidai (东信时代) was described as applying for a voice-driven digital human generation patent aimed at improving naturalness of performance, treating speech as the primary control signal for synthesizing facial and performance output; Yunxi Network Technology (Hebei) (云袭网络技术河北有限公司) was described as securing a patent for digital-human expression and facial-feature migration, focusing on transferring expressions and facial characteristics as a technical pathway to consistent identity and controllable performance; a personal-life “digital double” application was described through the launch of “Mever Space” (元我宙Mever空间), positioning user-created “virtual digital humans” as long-term personal archives built from diaries, life chronologies, and autobiographical writing, combined with dynamic family relationship graphs and an “AI genealogy” concept that extends the avatar from an individual memory container to a family lineage record.
Governance And Regulation: In Guangdong, ecommerce governance coverage described digital-human livestreaming content being brought within regulatory scope, treating synthetic hosts as accountable broadcast actors rather than novelty presenters and positioning enforcement attention on content compliance in avatar-driven live commerce; in insurance operations, Ping An Property & Casualty Insurance (China) (平安产险) was described as running a “digital human” project to improve sales development efficiency and service quality, framing the avatar as a controlled, auditable service-and-sales interface within regulated financial services rather than an unconstrained consumer chatbot.
February 04 News
Healthcare: Xiangya Jiangxi Hospital (湘雅江西医院) deployed an AI digital human named “Xiaoyaxi” (小雅西) in an intelligent guidance self-service kiosk to support on-site navigation and patient-facing inquiry handling; in Zhejiang, an “AI + healthcare” implementation described a unified “Health Brain” (健康大脑) approach that integrates cross-hospital medical information into comprehensive resident health records and exposes a digital-human health service system to residents as an access layer for individualized health services.
Marketing & Commerce: Soul App (Soul) through its Soul AI Lab (Soul AI Lab) open-sourced the real-time digital human generation model SoulX-FlashTalk (SoulX-FlashTalk), described as supporting sub-second end-to-end latency around 0.87s, 32fps high frame rate output, and stable long-duration video generation, positioning it for low-latency live interactive use; in Shaanxi reporting on apparel workflows, a “digital human helper” was described as enabling 24-hour uninterrupted livestreaming to support marketing and cross-border live commerce needs, reframing livestream operations as a continuous, automated presenter layer rather than a human-only staffing model.
Public Services & Government: Mofa Technology (魔珐科技) was described as building government-service digital human kiosks and integrated setups, with the AI digital human “Xu Xiaoxuan” (徐小宣) positioned as a district-level civic promotion identity and a digital host within the Shanghai Xuhui media matrix, spanning a full chain from publicity to service delivery; separate reporting highlighted public-facing “smart digital humans” in civic cultural spaces, including “Tu Xiao’an” (图小安) in the Xiong’an Library hall as an on-site interactive guide and digital-human interaction features embedded into a Shanghai Songjiang tourism comprehensive service center to attract and retain visitor engagement through humanlike conversational touchpoints.
Culture & Tourism: A Liaoning museum case described a virtual digital human “Liao Xiaobo” (辽小博), modeled on the collection artifact “Jade Pig-Dragon” (玉猪龙), operating as an AI-guided cultural companion that shifted visitor service from one-way explanations to two-way question-answer dialogue and accumulated over 100,000 answered audience questions after launch; a regional tourism integration example described the Guizhou cultural-tourism digital human “Huang Xiaoxi” (黄小西) being connected to the “One-Code Tour Guizhou” platform (一码游贵州) to aggregate provincial tourism resources and provide a single-entry guidance layer across attractions and services.
Media & Entertainment: Huiboxing (慧播星) was referenced as a digital human used in the production pipeline for video content distributed through mainstream portals, indicating routine substitution of a synthetic presenter for narration or hosting roles in news-style and social-content contexts; in Changsha County, digital human technology was applied to an AI micro–short drama format to retell the life story of Yang Kaihui (杨开慧) in a style framed as appealing to younger audiences, treating the digital human not as a standalone “anchor” but as a narrative performance device for historical storytelling.
Enterprise & Industrial: Haier Consumer Finance (海尔消费金融) filed a patent for a full-process, intelligent interactive credit system that includes a 3D digital human credit assistant, specifying an expression system based on FACS and speech synthesis based on Tacotron2 to deliver facial expression and voice outputs as part of the interaction loop; China Mobile Migu (咪咕) through Migu New Space Culture Technology (Xiamen) (咪咕新空文化科技（厦门）有限公司) filed a patent for a virtual human action generation method intended to reduce the production cost of virtual human motion; Fengyuzhu (风语筑) was described in capital-market coverage as using Baidu (百度) ERNIE Bot (文心一言) to provide capability support for AI-driven digital humans with multimodal interaction in its virtual human business scenarios, indicating an upstream large-model layer being positioned as the enabling substrate for digital-human perception and response behaviors in deployed applications.
Governance & Regulation: Multiple policy digests tied to February implementation timelines described regulatory measures for livestream e-commerce that explicitly bring AI-generated content such as “digital human anchors” into scope, treating synthetic presenters as regulated content actors within livestream commercial practices and signaling compliance expectations around the disclosure, management, and oversight of AI-generated presenting and promotional content in sales-oriented streams. (State Administration for Market Regulation (国家市场监督管理总局), together with Cyberspace Administration of China (国家互联网信息办公室), jointly issued livestream e-commerce supervision measures that explicitly bring AI-generated presenters such as “digital human anchors” into regulated scope, effective 1 February 2026, with compliance expectations around labeling/disclosure and the management and oversight of AI-generated promotional content in sales streams.)
February 03 News
Enterprise And Regulation: Several China-based policy and compliance developments explicitly brought “digital human” and “virtual human” outputs into platform-governance and live-commerce enforcement scope from February 1, 2026, treating digital-human livestream hosts as regulated “generated content” and linking compliance to platform rules, traffic/visibility controls, and live-commerce supervision mechanisms described in policy explainers circulated in Shaanxi Cyberspace Administration (网信陕西)–affiliated reporting and parallel coverage across multiple local outlets; in a related identity-and-rights direction, Beijing Economic-Technological Development Area (北京经济技术开发区) was reported as issuing what was described as a first-of-its-kind virtual-idol identity authentication, framing it as an institutional mechanism to clarify legal/social identity for virtual beings and to reduce friction around commercial exploitation, ownership confirmation, and rights protection for virtual idol IP and adjacent “digital employee” use in business contexts.
Education And Museums: A concrete classroom-and-museum case in Nanjing described an “AI digital human” demonstrating movable-type printing procedures inside a printing-culture museum setting, positioning the digital human as an embodied explainer for craft-process education rather than a generic chatbot interface, with the pedagogical function centered on step-by-step procedural demonstration for students and visitors. (The “AI digital human Bi Sheng” presentation is attributed to the museum’s builder and operator Jiangsu Phoenix Xinhua Printing Group (江苏凤凰新华印务集团有限公司), with the same coverage also stating that Modern Express (现代快报) applied its AI production capabilities in related museum-facing content featuring the Bi Sheng digital human.)
Tourism And Cultural Heritage: Multiple tourism deployments emphasized digital humans as on-site or platform-embedded explainers, companions, and interactive Q&A agents, including a reported “virtual human guide” pattern framed as a companion-like tour assistant and paired with “time-space reconstruction”/immersive historical-scene re-creation to let visitors experience reconstructed past environments; specific case studies included a Luzhou liquor-culture museum digital human named “Hanhan (酣酣)” demonstrating “ancient brewing” through MR mixed reality with visitors viewing holographic-style content via wearable or smart viewing devices, an “Ili General (伊犁将军)” digital human used as an explainer and question-answering role within an MR-enhanced historical overlay experience (including interactive story beats such as patrol scenes and city restoration), and a Guizhou tourism digital human “Huang Xiaoxi (黄小西)” integrated into the “One-Code Tour Guizhou (一码游贵州)” platform to aggregate provincial scenic-area information and provide unified consultation/interaction inside a single regional travel service entry point; destination-side corporate exploration also appeared via Guilin Tourism (桂林旅游) being reported as actively tracking AI and exploring how “virtual digital humans” could be applied to tourism services, while an urban tourism-service-center launch in Songjiang referenced “cultural-tourism digital human interaction” as part of an on-site service and local-products engagement setup.
Entertainment And Media IP: A notable entertainment case described an ultra-realistic digital human “Teng Xiaoma (腾小马)” reportedly modeled on a well-known comedian’s likeness and discussed as a prospective Spring Festival Gala-style performance partner for a human performer, framing the virtual being as a co-performer built around recognizable IP and photoreal character rendering rather than a simple animated mascot; in parallel, miHoYo (米哈游) was described as building an internal deep-learning R&D line (from an “Inverse Entropy” team to a formal institute) that produced a virtual digital human “Luming (鹿鸣),” positioning the character as an outcome of applied AI research connected to game-content production pipelines and broader synthetic-character capabilities.
Marketing And Commerce: Commercial-scale application was highlighted through JD.com (京东) being described as pushing AI-native digital humans into large-scale holiday retail scenarios as part of an ongoing commercialization track, while Tiandi Online (天地在线) was described as offering an AI marketing service chain that includes AI digital humans and XR livestreaming, associating digital humans with end-to-end campaign execution capabilities spanning virtual IP operations and immersive live-commerce presentation formats.
Public Services And Governance: A public-safety communications case described MoFa Technology (魔珐科技) producing a government-facing “digital human all-in-one” delivery and, in a concrete deployment, supporting AI digital human police personas “Qingfeng (清风)” and “Mingyue (明月)” for the Hangzhou public-security context, where the virtual beings were used to rapidly generate and continuously update short anti-fraud videos explaining scam patterns and prevention tactics as an ongoing public-information series rather than a one-off character reveal.
AI Companions And Personal Agents: A consumer social-product case described an AI-agent “digital double” application, A-Quan (啊圈), launched on iOS with the premise of “intelligent avatars” acting as personal agents for social participation and application execution, while an enterprise training case described Nengshi Gaotu AI (能师高徒AI) working with Taoyu Consulting (韬钰咨询) on a “doppelganger intelligent-agent platform” and an “AI mentor solution” that structurally extracts role-expert tacit knowledge, judgment patterns, and working methods, then performs semantic modeling and training so the resulting agent behaves as a job-aligned expert proxy in real business-post scenarios.
Platforms And Tooling: Toolchain progress included Soul App (Soul) announcing the open-sourcing of a real-time digital-human generation model, “SoulX-FlashTalk,” explicitly positioning the release around real-time talking-head/digital-human generation capability as reusable infrastructure for synthetic-character production and interactive avatar experiences rather than a closed, single-application feature.
February 02 News
Enterprise: Jiufang Zhitou (九方智投) referenced its in-house “Jiufang digital human” as part of an “AI+manufacturing” framing that also listed intelligent image recognition, AIGC intelligent content creation, and emotion/sentiment tendency analysis as related capability areas, and it linked this manufacturing-oriented AI agenda to the national AI application pilot-base effort for the manufacturing domain that is jointly guided by National Development and Reform Commission (国家发展改革委) and Ministry of Industry and Information Technology (工业和信息化部) and constructed by Shanghai Electric (上海电气).
Tourism: In Shanghai (上海), Jiefang Daily (解放日报) described a new Songjiang all-region tourism integrated service center launch that included on-site “cultural-tourism digital human” interactive installations as a visitor-facing experience element; separately, Metaverse Matrix (元界矩阵) announced an AI companion touring “sprite” agent called “Miaoyuanzhang” (喵元章) deployed via an on-site “sprite cabinet” installation at the Chongqing Garden Expo Park Spring Festival lantern-fair event window (January 31 to March 22), positioning the agent as an AI companion for guided visiting in that park-based festival setting.
Cultural: Xinhua News Agency (新华网上海频道) cited an AIGC microfilm titled “Password in the Notes” (《音符中的密码》) that used an “AI historical scene reconstruction system” together with a digital-human component as part of a red-culture digital dissemination approach, and it associated the related performing-arts production context with Impression Honggu (Shanghai) Performing Arts Co., Ltd. (印象红谷（上海）演艺有限公司) as an affiliate of Shanghai Zhonggong Muhua Enterprise Service Group (上海中工沐华企业服务集团); in another concrete heritage-activation case, a Shanxi Museum project was described as creating a digital human named “Qingniao” (青鸟) modeled on the collection piece “Jinhou Bird Zun” (晋侯鸟尊), using holographic projection to “animate” the artifact and pairing it with a customized 12-episode cultural-tourism interpretation/explainer content line.
Education: A “smart education” deployment example was described as a full-modal interactive digital human used in school settings, with landing units named as Chengdu Tianfu Middle School and Chengdu Gaoxin Yunxin School, presenting the digital human as the interaction layer for education scenarios rather than as a purely decorative avatar.
Marketing And Media: A Spring Festival Gala example described “Doubao” (豆包) appearing through a virtual digital-human format that conducted real-time dialogue and interstitial interaction with human hosts, and it also described user-facing creation where natural-language instructions could generate personalized outputs tied to that same “participatory partner” interaction framing; in video-generation tooling, Kunlun Wanwei (昆仑万维) announced an open-source multimodal video generation model under its Skywork AI line called SkyReels-V3, explicitly listing “audio-driven virtual avatar” capability alongside reference-image-to-video and video extension, positioning the virtual-avatar function as an input modality for controllable character/host-style outputs in generated video.
Governance: Multiple notices about new live-ecommerce oversight rules described digital-human anchors as a regulated form of AI-generated content in livestream commerce, including requirements that AI-generated content such as digital-human hosting be clearly marked and that livestream operators bear responsibility for prohibited outcomes such as false promotion and vulgar content, framing “digital human anchor” compliance as a concrete operational requirement for platforms and merchants using synthetic presenters.
Enabling Technologies: Daoyoudao Technology Group (道有道科技集团股份公司) filed a patent application for a modular-parameter-based virtual digital human generation method and system, explicitly emphasizing automatic generation without manual work and aiming for more natural animation behavior; separately, Sizhi Era Technology (广州思智时代科技有限公司) filed a patent application for an interaction rendering and testing method for virtual digital humans, explicitly targeting improved facial-expression rendering accuracy/precision as a quality metric for interactive digital-human presentation.
February 01 News
Healthcare: Aier Eye Hospital (爱尔眼科) was reported as having a “digital human foundation development and application” project selected as one of Hunan Province’s “digital new infrastructure” landmark projects, framed as a step toward deeper integration of artificial intelligence with clinical practice by leveraging very large-scale medical data resources and positioning the digital-human layer as an enabling base for future medical-facing applications. (Aier Eye Hospital Group (爱尔眼科医院集团) is described as co-developing with the Institute of Computing Technology, Chinese Academy of Sciences (中国科学院计算技术研究所) an ophthalmology-focused large-language-model system and associated “AierGPT” plus “Aier digital human” outputs, which were publicly showcased in 2025 and framed as a digital-human foundation layer intended to support AI integration into clinical ophthalmic practice.)
Education: In Hunan, a province-level immersion-style AI teaching workshop explicitly included digital-human development alongside AI audio-visual creation, intelligent office workflows, low-code intelligent-agent customization, and local large-model deployment, indicating that “building digital humans” is being treated as a practical educator skill set rather than a purely research topic; in Sichuan (Lezhi), an “AI digital human + mathematics game” teaching design was cited as having won a top prize in a higher-level teaching competition, presented as a concrete classroom implementation of a digital-human-mediated instructional approach.
Tourism And Cultural Heritage: In Jiangsu (Nanjing), cultural-sector coverage described the emergence of “digital human + livestream” and “AI + culture” as named new formats, tied to cultural programming and to packaging large heritage narratives such as the China Grand Canal into digital-first experiences; in Shanxi, Shiyou Technology (世优科技) was described as building a Chinese-style digital human “Qingniao” (青鸟) based on the Shanxi Museum’s signature artifact “Jin Hou Bird Zun” (晋侯鸟尊), with Qingniao deployed as a digital host at a tourism development conference and positioned as a heritage-interpretation and event-presenting character enabled by virtual–real fusion techniques; in Tianjin, the National Maritime Museum of China (国家海洋博物馆) “Future Ocean” hall introduced an AI digital human “Lingxi” (灵汐) as part of an interactive exhibition design intended to shift visitors from passive viewing to guided, participatory interaction; in Qinghai (Xining), reporting on a “digital-intelligence commercial district” explicitly referenced “digital-human shopping guides” used alongside AI-guided shopping assistants, presenting the digital-human role as a front-end retail concierge within a broader smart-consumption environment.
Media And Entertainment: China International Television Corporation (中视传媒) stated that the CCTV virtual host “Xiao8” (小8) was produced by CCTV Northern Production (中视北方) and emphasized that it did not use AI modeling or motion capture and lacked digital interactive support, which frames Xiao8 as a virtual-presenter case grounded in conventional production rather than AI-driven digital-human pipelines; Beijing International Film Festival AIGC Film Unit (北京国际电影节AIGC电影单元) solicitation materials highlighted institutional participation that includes China Film Artificial Intelligence Research Institute (中影人工智能研究院), Beijing Metaverse Sci-Fi Future Technology Research Institute (北京元宇科幻未来技术研究院), and Communication University of China Digital Human Research Institute (中国传媒大学数字人研究院), signalling that digital-human research bodies are being positioned as part of the festival’s AIGC film ecosystem and evaluation context; separately, cultural reporting referenced virtual-human vocal technology being used to give traditional-instrument narratives a “digital life” within a branded program context, presented as a content-production technique for re-staging heritage music through synthetic vocal performance.
Marketing And Commerce: China Telecom (中国电信) was reported as operating an “AI + Xingboke omnichannel livestream matrix platform” that directly deploys AI digital-human anchors to run a “never-ending livestream room,” presenting the digital human as an always-on host for couponing and conversion flows without dependence on "human anchor" scheduling; in Gansu (Lanzhou), coverage of a “beef-noodle metaverse experience hall” described a consumer loop combining digital-human interactive ordering, production visualization, online check-ins, and offline redemption, with the digital human functioning as the interactive front-end for ordering and engagement inside a hybrid online–offline retail experience.
Public Service And Civic Venues: Reporting on a youth legal-education exhibition venue in Tianjin described a digital-human interaction system that uses an ultra-realistic digital human to provide face-to-face services such as consultation and reminders, positioning the virtual being as an on-site public-service interface designed to feel more immediate and approachable than conventional signage or kiosks; in Xinjiang, a memorial-style exhibition note stated that a video segment was “made using Huiboxing digital human” (慧播星数字人), indicating a concrete tool-based workflow where a packaged digital-human product is used to generate or present narrative video content for public communication.
Enterprise And Finance: Industrial and Commercial Bank of China (工商银行) was reported as receiving authorization for an invention patent describing an information-interaction method that presents response videos to users in the form of a virtual digital human, framing the virtual being as a standardized, bank-controlled response layer for customer-facing information delivery rather than as a one-off marketing avatar. (Industrial and Commercial Bank of China discloses in its “Rong e Xing” (融e行) privacy policy that its digital-human service uses a digital-human SDK provided by Beijing SenseTime Technology Development Co., Ltd.)
Technology And R&D: Kunlun Wanwei (昆仑万维) was reported as open-sourcing SkyReels-V3, described as a unified multimodal video-generation system that supports reference-image-to-video generation, video continuation/extension, and audio-driven virtual avatar/digital-human generation within a single architecture, with the audio-driven capability explicitly tied to producing a speaking or performing virtual character from portrait reference and sound; ByteDance (字节跳动) was reported as releasing FlowAct-R1 as a real-time interactive virtual digital-human advance, with accompanying acknowledgement that higher-fidelity virtual-human interaction can be misused for harmful or deceptive content and that the team planned access controls as part of responsible deployment; Guangzhou Sizhi Shidai Technology (广州思智时代科技有限公司) was reported as filing a patent application for a virtual digital-human interactive rendering and testing method aimed at improving facial-expression rendering accuracy, positioning expression fidelity as a measurable engineering target supported by test-oriented tooling.
Governance, Standards, And Regulation: The Ministry of Industry and Information Technology (工业和信息化部) was reported as publicly soliciting opinions on a mandatory national-standard project titled “Metaverse classification and labeling—digital human identity identification requirements,” which frames “digital human identity labeling” as a regulated interoperability and compliance requirement within metaverse categorization and marking systems; local “new rules” coverage in Shanxi (Taiyuan) and elsewhere described livestream e-commerce governance proposals that explicitly bring “digital-human anchors” and other AI-generated content into supervision frameworks by assigning legal responsibility for violations to the operators or users of the livestream rooms that deploy such generated content; Beijing governance coverage also referenced disputes over virtual-digital-human rights and platform liability in the context of infringement claims, presenting virtual-human assets as protectable creative works and positioning platform participation in distribution as a factor in shared responsibility.