June 2025 News

June 30 News

Marketing & E-Commerce: In China's e-commerce sector, digital humans are operating as active commercial agents in high-stakes livestream environments. The digital human Shuibeibei was deployed as a central presenter during the 2025 Shuibei Gold & Jewelry E-Commerce Livestream Season in Shenzhen, an event that generated over 8 billion yuan in sales across 40 days, demonstrating the commercial viability of synthetic characters in consumer retail. Zhejiang Mobile (浙江移动) has developed a multi-language product promotion tool that uses AI-generated digital humans to convert a brief spoken prompt into broadcast-ready multilingual video content, enabling rapid international marketing at scale. Both developments were highlighted at the 2025 AIGC and Smart Design Innovation Forum in Hangzhou, signaling institutional recognition of digital humans as core infrastructure in China's digital commerce ecosystem.

Enterprise Communications: Digital humans are increasingly adopted in China for corporate communications and strategic thought leadership. 360 Group (三六零集团) founder Zhou Hongyi appeared at the Zhenghe Island 2025 Case Learning Annual Meeting not in person but as an AI digital human, delivering a keynote titled "Development Trends and Practice Paths of Intelligent Agents," an application that positions synthetic avatars as credible stand-ins for senior technology executives in formal business contexts. In parallel, Jisi Technology, operating within Beijing's Economic-Technological Development Area at the National Information Innovation Park, has developed a streamlined production pipeline capable of generating a broadcast-ready AI digital human from real-person video and audio capture in approximately two hours, reflecting the industrialization of synthetic media workflows for public-facing content deployment.

Technology Development: At the platform and infrastructure level, Tencent (腾讯) has reported significant advances in digital human development, releasing related code and models publicly. This shift toward open access represents a meaningful change in how enterprise-level digital human production is structured in China, with the potential to lower barriers for downstream developers and businesses seeking to integrate lifelike synthetic characters into products and services at scale.

June 29 News

Enterprise: Meituan (美团) has deployed AI digital humans as virtual employees within its service retail operations, positioning synthetic characters as functional staff capable of handling customer interactions and improving operational efficiency in online commercial environments. Baidu (百度) unveiled a dual-digital-human live broadcasting system driven by its Wenxin 4.5 Turbo large multimodal model, demonstrating synchronized virtual host interactions designed for commercial live streaming applications. Together these deployments represent a concrete shift among major Chinese internet and retail platforms toward integrating digital humans into core service and sales workflows rather than treating them as supplementary features.

Technology: Inspur Smart City Technology Co., Ltd. (浪潮智慧城市科技有限公司) filed patent CN120216646A covering a digital human-based intelligent dialogue system, signaling the company's intent to advance naturalistic, human-like AI interaction capabilities within enterprise and urban service contexts. The patent reflects a broader drive within China's technology sector to embed virtual human interfaces into smart city infrastructure, extending the functional reach of digital humans beyond entertainment and into civic and administrative service delivery.

Media: Two research papers — "Research on the Role of Generative AI in China's Media Landscape" and "Investigation of the Function of Generative AI in China's Media Environment," both by S. Hao and M. Sisugoswami — document how AI-generated content, including synthetic characters functioning as virtual anchors and AI-driven news presenters, is restructuring China's media ecology. These studies characterize the proliferation of virtual hosts in broadcast and online news contexts as integral to state-aligned modernization strategies, with generative AI enabling a reconfiguration of traditional media roles through the deployment of digital human presenters at scale.

Marketing: An empirical study focused on live-streaming in China's tourism and hospitality e-commerce sector examines how virtual influencers operating as live-streaming hosts influence consumer purchasing decisions. The research identifies trust, emotional engagement, and authenticity perception as key mediators of audience stickiness and commercial intent, situating virtual beings not merely as content vehicles but as active commercial actors capable of driving conversion within China's livestream retail ecosystem.

Cultural: The virtual digital human Xiao Bo (小博) was deployed at the China-Singapore Friendship Library to deliver interactive cultural and educational experiences, combining physical spatial navigation by robot with digital storytelling through a virtual being interface. This implementation illustrates how virtual humans are being applied in Sino-international institutional settings to support bilingual and cross-cultural engagement, extending the functional presence of digital humans into public library and heritage environments.

June 28 News

Healthcare: Ant Group (蚂蚁集团) has developed an AI-powered healthcare application that presents users with AI avatars representing more than 200 prominent Chinese doctors, enabling interactive medical consultations and patient support through synthetic character interfaces. This deployment positions virtual beings as a functional layer in healthcare access, combining the recognizable authority of real practitioners with the scalability of AI-driven avatar systems.

Marketing and Commerce: JD.com (京东) reported that its Yanxi digital human service platform supported over 9,000 brands during the "618" shopping festival, generating more than 400,000 hours of livestreaming content and attracting over 100 million viewers, demonstrating the scale at which AI digital humans are operating within China's e-commerce environment. Baidu (百度) deployed digital humans in collaborative livestreams including a dual-avatar setup replicating tech influencer Luo Yonghao; a one-to-one digital likeness of Luo achieved 13 million views and 55 million RMB in gross merchandise volume, illustrating the commercial effectiveness of celebrity-replicated virtual beings. Ranmai Technology (燃麦科技) continues to deploy its virtual influencer Ayayi across platforms including Xiaohongshu, where she appears in livestream-style short videos promoting brands and engaging audiences. The virtual character 红鲤FishRed, oriented toward Gen Z engagement, has appeared in brand interactions, while 柳夜熙 (Liu Yexi), a digital human and short-form content creator, has built a significant following on Douyin through alignment with trending visual aesthetics. Zoomlion (中联重科) commissioned the creation of a digital spokesperson named Xiao Wei (小薇), modeled on a Younger Sister (幺妹儿) archetype, to mark the company's 30th anniversary; the project was led by CEO Wang Tianxiao of Hunan Glass Culture (湖南玻璃文化) and was framed as a merging of digital marketing with advanced technology for brand storytelling. Chinese-language media also reported on virtual try-on systems allowing users to preview clothing from various sources — including vintage shops and social media trends — using AI-generated virtual likenesses, reflecting the diffusion of avatar-based retail tools into broader consumer culture.

Enterprise and Governance: China Telecom (中国电信) demonstrated multiple enterprise-facing virtual being applications at the Quanzhou Cultural Tourism Conference, including an AI digital human replication system, an AI-driven tourism recommender, and a holiday traffic prediction platform, integrating avatar technology into civic service infrastructure and institutional decision-making. The digital worker Xiao Yang (小漾), deployed by Shanghai's Yangpu District government, represents a broader trend in which public-sector entities adopt virtual beings for administrative, promotional, and interactional roles. The virtual anchor New Little Raccoon (新小浣熊君), associated with Weishi (微视), has also been cited in the context of digital governance and customer-facing institutional communication, reflecting how AI-generated characters are being embedded across both government and enterprise service delivery.

Entertainment: The hyper-realistic digital human Kong Linghui (超写实数字人) blends the public profile of a traditional sports figure with digital reinterpretation, functioning as both entertainer and virtual celebrity. HaoFeiSi Technology backed the fictional character Vivian as a tech-infused cultural ambassador blending AI aesthetics with pop culture to broaden public engagement with digital human figures. Multiple digital idols and virtual stars have been integrated into short dramas and livestream events, occupying roles in entertainment formats traditionally filled by human performers and reinforcing the growing currency of synthetic characters within China's content economy.

Technology and Production: Tencent (腾讯) released HunyuanVideo-Avatar, a system employing advanced machine learning to generate lifelike digital avatars capable of mimicking human expressions, movements, and voices with high accuracy, positioned for deployment across entertainment and professional digital platforms. ZEGOCLOUD (即构科技) offers a PaaS SDK enabling developers to generate AI digital humans for video and live-streaming applications, supporting use cases including short video production, AI companionship, customer service, and real-time interactive digital human livestreaming. UBTech (优必选) filed a patent for facial motion capture optimized for 2D virtual humans, signaling continued technical investment in avatar expression and realism. Chinese productions associated with director Guo Jingming, building on projects following L.O.R.D: Legend of Ravaging Dynasties, have highlighted the complexity of integrating XR technology with digital humans, requiring detailed pre-production and filming adjustments to maintain realism and depth perception in narrative media. At the 2025 Summer Davos Forum in China, companies demonstrated virtual human systems enabling face-to-face interaction between users and AI-driven characters through integrated voice recognition, semantic understanding, and speech synthesis, with the technology framed as a growth engine for immersive communication sectors.

Education: Henan Light Industry Vocational College (河南轻工职业学院) launched a competitive procurement project for an AI illustration and virtual digital human training program, designed to cultivate applied skills in creating and interacting with AI-generated human-like figures. This initiative positions virtual humans as practical vocational tools and reflects growing institutional interest in equipping students with competencies in synthetic media and avatar-based interfaces within China's creative and technology industries.

Cultural Preservation and Tourism: Virtual characters in China are increasingly applied to cultural heritage and regional tourism contexts. The character Wen Yaoyao (文夭夭) serves as a virtual museum guide and narrator of traditional stories, bridging digital content with heritage preservation through synthetic character interfaces. Yao Xiaoman (瑶小蛮) functions as a digital host tied to Hunan ethnic culture, supporting regional identity and tourism promotion through avatar-based storytelling and reflecting how digital humans are being tailored for culturally specific and geographically grounded communication purposes.

Legal and Regulatory: China's virtual influencer economy has generated notable legal developments around the concept of 中之人 (zhōng zhī rén), referring to the real individuals who animate virtual personas. A breach of contract case led courts to recognize the reuse value of virtual characters and to incorporate that value into damage calculations, marking a step toward more sophisticated legal frameworks for synthetic identity and brand control. These developments coincide with IDC projections estimating the AI digital human market in China at approximately 4.12 billion RMB by 2024, underscoring the pace of commercial expansion that is driving both regulatory attention and the broader legal reckoning with synthetic character ownership.

June 27 News

E-commerce: Baidu invested 100 million RMB in subsidies to promote digital humans as livestream hosts under its "Mengdie" and "Fanxing" plans, with a stated goal of creating 100,000 digital human livestreaming sales influencers. This initiative represents one of the most explicitly scaled deployments of synthetic characters in China's live-commerce sector, positioning AI-driven virtual beings as a systematic replacement or supplement to human hosts in sales-oriented broadcast environments.

Enterprise: Alibaba's "TaoAvatar," developed by the Taobao Meta technology team, enables real-time rendering of 3D digital humans and supports AI-driven conversations on smartphones and XR devices, and has been released as an open-source solution to facilitate broad commercial adoption. At a technology forum associated with the 2025 Summer Davos Forum, a company marketing manager demonstrated how users can engage in face-to-face-style communication with virtual humans through integrated speech recognition, semantic understanding, and voice synthesis, illustrating how Chinese enterprises are deploying virtual beings in personalized customer engagement and marketing contexts. Mingtou Technology's platform "Zhixiaotu" (智小途), introduced by Market Director Li Hui, enables individuals to create digital human clones built on personalized small-model knowledge bases tailored to specific job functions, with these avatars serving as intelligent agents in new professional roles by combining gaming mechanics with workplace productivity tools.

Entertainment: Huanrui Century (欢瑞世纪) was reported among Chinese stocks experiencing gains directly attributed to the company's involvement in short dramas and interactive games featuring virtual digital humans, reflecting the commercial traction of synthetic characters in immersive storytelling and interactive media. Dongfang Guoxin and Huanyu Printing saw comparable market movement linked to virtual digital human sector interest, with both companies engaged in technologies including blockchain and short drama production that intersect with digital human deployment in entertainment contexts.

June 26 News

E-Commerce and Livestreaming: A digital human livestream in China generated sales of 55 million yuan, as reported by 华盛通, prompting public debate about whether AI-powered virtual presenters could supplant prominent human influencers such as Luo Yonghao. Kuaishou (快手) has taken an active institutional stance in this space by promoting AIGC digital humans across its platform, signaling strong corporate endorsement of synthetic content delivery for commercial purposes. Alibaba's (阿里巴巴) Taobao Meta team entered this competition with the release of TaoAvatar, a 3D digital human system optimized for mobile deployment that supports real-time AI dialogue and high-fidelity rendering, positioning it for use in interactive e-commerce and consumer-facing livestreaming contexts.

Technology and Platforms: TaoAvatar employs Gaussian splatting to achieve high-visual-fidelity rendering on mobile and XR devices, representing a notable technical advance in making production-quality digital humans accessible outside enterprise hardware environments. Grey Dolphin (灰豚) has released open-source tooling for creating highly realistic digital humans with integrated beautification capabilities, contributing to a growing ecosystem of accessible development resources. A technical analysis of the infrastructure supporting digital human deployment in China identifies significant hardware and software dependencies required to sustain scalable and synchronized virtual human interactions, particularly within enterprise and public-facing platforms.

Cultural Applications: In Shanxi, an AR-enhanced cultural tourism experience features a virtual representation of Tang Dynasty poet Wang Wei emerging from ink paintings to recite poetry, merging heritage education with avatar-based embodiment to offer visitors an immersive interpretive layer for historical content. Chinese funerary practices have also incorporated digital human technology, with reports from 하늘문화신문 documenting the use of AI-generated digital resurrections to memorialize deceased individuals, reflecting a culturally significant application of virtual being systems in mourning and memory. The Chinese AI digital human industry has further institutionalized three distinct roles for virtual idols — virtual singers, virtual UP主 (content uploaders), and virtual bloggers — positioning these synthetic identities as central figures within a maturing content ecosystem that integrates entertainment, AI, and user engagement.

Legal and Ethical Contexts: A landmark civil case in Suzhou examined the contractual breach of a virtual anchor's human operator, referred to as 中之人, raising foundational questions about labor rights, legal accountability, and intellectual property in the virtual influencer industry where human performance and digital representation are deeply intertwined. A case reported by 科技日报 describes a courtroom scenario in which a virtual character emotionally articulated a belief in forgiveness during simulated legal proceedings, with the presiding judge accepting this as testimony, introducing serious ethical questions about authenticity and the potential for emotional manipulation through AI-generated legal narratives. Separately, Sina Finance reported on criminal exploitation of AI face-swapping technology using virtual digital human simulations to commit fraud, prompting law enforcement to conduct public awareness campaigns, with particular focus on educating elderly populations about the risks of synthetic identity deception.

Enterprise and Public Sector: During China's 2025 National Low Carbon Day event held at Hangzhou Future Sci-Tech City, an AI virtual host guided participants through programming centered on ecological transformation, demonstrating how government-backed institutional events are beginning to deploy virtual humans as authoritative and engaging public-facing presenters.

June 25 News

E-Commerce and Livestreaming: Baidu (百度) partnered with internet personality Luo Yonghao to deploy an AI-generated digital human version of him for livestream e-commerce, with the virtual replica hosting its debut session on Baidu Youxuan and drawing 13 million viewers while generating 55 million RMB (approximately USD 10.5 million) in gross merchandise value, surpassing the performance of comparable human-led livestreams. In a subsequent seven-hour session during the 618 shopping festival, the digital human versions of both Luo Yonghao and presenter Zhu Xiaomu headlined the broadcast, achieving faster and higher sales conversions than their human counterparts. By 2023, digital human-led commerce had consolidated as a key driver of market scale expansion within China's virtual being economy, and the Baidu deployments are widely cited as a proof-of-concept for AI digital humans as scalable replacements for human influencers in high-volume retail events. Academic analysis published in 2025 by S. Mao, examining the evolution of business models across China's online live streaming platforms, details how synthetic characters and virtual hosts have become embedded in platform architecture, employed to maintain constant engagement, reduce human labor costs, and feed into recommendation and monetization algorithms originally designed around human streamers.

Industry Landscape: A 2024 industry white paper by iResearch formally identified the AI digital human interaction sector as having matured from a conceptual phase into active commercial implementation, designating it an essential domain within human-computer interaction. The top three companies recognized for technological leadership in this sector in 2025 were Shiyou Technology (世优科技), Xiaoice (小冰科技), and iFLYTEK (科大讯飞), each credited with developing platforms and services that enable naturalistic interaction between AI avatars and users across enterprise applications. These three companies represent distinct approaches to digital human deployment, with Xiaoice's heritage in conversational AI, iFLYTEK's foundation in speech synthesis and multimodal interaction, and Shiyou Technology's positioning as a specialist digital human platform, collectively illustrating the range of technical pathways being pursued within China's AI avatar industry.

Media and Broadcasting: In Laibin, Guangxi, the AI virtual news anchor Qingyang (清扬) has been integrated into news reporting across multiple television program formats, presenting with realistic facial expressions, natural pacing, and tonal variation comparable to human broadcasters. This deployment reflects a broader adoption of AI-generated virtual presenters within China's traditional broadcast media sector, where synthetic characters reduce dependence on human anchors while sustaining the visual and vocal conventions of broadcast journalism.

Education: Jinan Vocational College in Shandong Province has deployed AI digital humans as interactive tools for student career preparation, enabling students to engage in simulated job interviews, receive structured feedback, and refine their responses across repeated sessions. This application positions virtual beings as personalized pedagogical agents within China's vocational education system, addressing the practical needs of graduates competing in a demanding employment environment.

Social Platforms: On Douyin (抖音), young users are increasingly mediating their social identities and interpersonal engagement through customized avatars and virtual characters rather than text or unmediated video, creating and personalizing digital personas to participate in shared visual experiences and community interaction. This shift marks a broader transition in Chinese social platform culture toward avatar-based self-expression, where virtual beings serve as primary interfaces for identity, connection, and entertainment rather than supplementary features layered onto existing formats.

June 24 News

Livestream Commerce: Digital avatars deployed on Baidu's (百度) e-commerce platform generated 32.8 million RMB in a single livestream session, substantially outperforming human influencers operating in comparable formats. These avatars demonstrated real-time expressive capabilities including gesture, speech, and live product demonstration, functioning as autonomous sales agents rather than passive visual assets. In a separate campaign, AI avatars operating in China's digital retail environment generated an estimated seven million USD in a single activation, further reinforcing the commercial potency of virtual beings in Chinese livestream commerce. Chinese influencers also deployed highly realistic avatars in livestream marketing operations targeting Thai audiences, a practice documented as early as April 2020 on short-video platforms, illustrating the cross-border extension of China's avatar-driven sales model.

Celebrity and Entertainment: Baidu launched a multimodal digital human modeled on Luo Yonghao, a prominent figure in China's technology and entertainment sectors, presenting it at Baidu AI Day as the industry's first high-fidelity multimodal digital host. The avatar was deployed on Baidu Youxuan (百度优选) for a livestream ecommerce event, achieving over 50 million RMB in gross merchandise value within four hours. The same digital human subsequently appeared in a Father's Day livestream campaign, demonstrating recurring commercial deployment of the avatar across distinct event contexts rather than one-off promotional use.

Technology and Production: Weimei Hologram (微美全息) has developed a production model enabling low-cost, short-cycle, batch creation of AI-driven interactive digital humans, with support for customized aesthetics to serve diverse platform and client requirements, accelerating deployment across media and broadcast applications. Chingmu updated its digital human livestream system to version 3, incorporating upgrades to both the voice and visual modules to bring avatar behavior and appearance closer to human norms in real-time broadcast environments. HeyGem, covered in LiuYue's Tech Blog, released a new digital human package with native 50-series GPU support, removing the requirement for Docker containers and adding batch inference and task management features, signaling a maturation of developer-facing digital human tooling in the Chinese technical ecosystem.

Market Scale and Enterprise Adoption: Industry estimates cited in Chinese financial media and IDC's 2024 assessment place the value of China's AI digital human market at 4.12 billion yuan for the year, reflecting an annual growth rate of 85.3% and indicating sustained and accelerating commercialization across live broadcasting, customer service, and brand representation. O-Net Communications (欧耐通信), a Guangdong-based telecom equipment manufacturer designated as a "专精特新" enterprise under China's specialized and innovative SME classification, has expanded into virtual digital human technologies alongside connected vehicle networks, representing the diversification of digital human investment beyond the media sector into industrial and enterprise applications. A separate reference to China's "National Avatar Strategy" signals coordinated policy-level interest in integrating virtual beings into both strategic communications and commercial infrastructure, framing digital human development as a matter of national industrial priority.

June 23 News

Commerce: During China's 618 shopping festival, digital humans made substantial commercial inroads across major Chinese e-commerce and livestreaming platforms. Taobao (淘宝) and Douyin (抖音) both recorded significant year-on-year increases in effective orders attributable to digital human deployment, at 65% and 42% respectively. On Baidu Youxuan (百度优选), two livestream sessions hosted by digital humans generated over 105 million yuan in gross merchandise value, marking a landmark moment for the industry. Separately, an AI avatar duo operating on Chinese video platforms was reported to have generated cumulative sales of 10.5 billion yuan, demonstrating the scale at which synthetic presenters can function as commercial assets in high-volume livestream environments.

Enterprise: China Central Television (中央电视台) and Sohu (搜狐) drew attention to emerging professional roles forming around digital human systems, identifying occupations such as AI digital human trainer and generative AI animation producer as new categories of enterprise labor. One profiled practitioner, Sun Jiaxin, works at a Beijing-based internet technology company employing nearly fifty staff dedicated to AI training functions, encompassing prompt design and visual model optimization. These roles reflect the integration of virtual beings into commercial production workflows, where human specialists are required to develop, refine, and maintain synthetic character systems for sustained deployment.

Healthcare: At a technology competition held in Xiong'an, the Chinese Academy of Medical Sciences' Fuwai Hospital (中国医学科学院阜外医院) presented a smart hospital solution built on digital twin technologies, incorporating virtual digital humans to simulate visual consultation scenarios and model a future architecture for hospital service delivery. The project was awarded first prize among 132 entries in the medical innovation category, signaling growing institutional recognition of virtual beings as functional instruments in clinical planning and healthcare service design.

Cultural: The Capital Library (首都图书馆) developed an AI digital human modeled on the renowned Chinese author Lu Xun, training deep learning models on his complete writings, associated research literature, and historical records, and deploying this virtual being at a cultural reading event along the Grand Canal in 2025. A separate cultural reference interprets jade dragons from approximately five thousand years ago as early symbolic avatars that accompanied ancient leaders, positioning them as evidence of a long-standing human tradition of constructing synthetic representations to mediate spiritual or social experience.

Research: A Chinese-language scientific study titled "Virtual Human Retina: Simulating Neural Signalling, Degeneration, and Responses to Electrical Stimulation" applies virtual human frameworks to biomedical modeling, using simulation-based methods to study retinal neural behavior, degeneration processes, and responses to electrical stimulation. The work positions virtual humans as controlled experimental environments within vision science, extending the application domain of synthetic human models into precision medical research.

June 22 News

Commerce: Baidu (百度) supplied the generative AI infrastructure behind multiple high-profile virtual human livestream commerce deployments in 2025. In Yuchuan, a Baidu-powered avatar duo featuring a co-host named Xiaomu generated approximately 10.5 billion won in sales within a single seven-hour broadcast, surpassing the performance of human influencers on the same platform. On June 15, serial entrepreneur Luo Yonghao launched a digital human avatar on Baidu's e-commerce platform, with the avatar's first broadcast producing a gross merchandise value exceeding one billion yen; Luo subsequently expressed surprise on Weibo that the avatar had outearned him during a comparable live appearance. In a separate event, Luo deployed twin AI avatars powered by Baidu's ERNIE model, drawing over 13 million viewers and generating more than ¥55 million in gross merchandise volume. JD.com (京东) similarly scaled virtual human deployment during the 618 shopping festival, positioning AI-generated digital employees as front-facing livestream hosts for major brands. These virtual hosts were engineered to operate continuously at high capacity, equipped with advanced sales logic, high emotional intelligence, meme generation capability, and crowd management functions, reflecting a move toward fully operationalized commercial deployment of digital humans in live e-commerce.

Technology and Media Production: Digital Domain (數字王國), the Hong Kong-based visual effects company, released an AI-powered video production solution constructed on its virtual human technology, positioning the system as a convergence of creative and technical capability intended to support ongoing short-form content production. Huawei Cloud (华为云) launched MetaStudio, an intelligent interactive digital human platform designed to make embodied AI assistants available to individual users, families, and enterprises, enabling the creation of personalized AI companions with interactive capabilities applicable across multiple sectors. Together, these releases illustrate how established technology companies in the Chinese-language market are moving virtual human infrastructure beyond experimental contexts and into scalable production and consumer-facing deployment frameworks.

Education and Cultural Heritage: Tianjin University of Finance and Economics initiated a program using AIGC technologies to develop AI-generated virtual digital humans serving as departmental ambassadors, designed to present immersive educational scenarios and function as dynamic interfaces for illustrating professional capability frameworks across academic disciplines. In Wuxiang, a virtual digital human named Red Star Yang, dressed in Eighth Route Army attire, was deployed as part of a historically themed immersive experience supported by 5G cloud rendering and metaverse reconstruction of wartime environments, serving more than 300 visitors daily. Both cases demonstrate the application of virtual human technology in institutional and heritage contexts, where the priority is experiential engagement rather than commercial performance.

Companion Applications: The Dudou APP frames its AI digital humans as emotionally aware companions rather than functional interfaces or sales agents, promoting them explicitly as "soulmates" for users and emphasizing personalized interaction and emotional resonance as primary design goals. This positions the product within a growing segment of the Chinese digital human market oriented toward sustained affective engagement, where the commercial proposition depends on depth of emotional connection rather than transactional utility.

Platform Governance: Thai-language reporting on Chinese virtual human deployments in livestream commerce identified platform compliance, rather than technological capability, as the principal barrier to broader adoption. The analysis noted that digital humans must be trained to conform to platform-specific regulations governing speech, behavior, and user engagement, representing a shift in industry focus from development challenges to governance and operational integrity concerns.

Marketing Research: Research by W. Zhu and colleagues examined how influencer traits on Douyin, including trustworthiness and expertise, shape the purchase intentions of electric vehicle consumers in Nanjing, providing quantitative grounding for the strategic design of virtual and human influencer personas deployed on Chinese short-video platforms.

June 21 News

Marketing and E-commerce: The deployment of AI-generated digital humans in China's livestream e-commerce sector has produced commercially significant results. Baidu (百度) developed a generative AI avatar trained on five years of video content to replicate the humor, speech patterns, and on-screen presence of prominent host Luo Yonghao. During a single livestream event, this avatar co-hosted alongside a digital replica of co-host Xiao Mu, generating a gross merchandise value of 225 million yuan over seven hours and attracting more than 13 million viewers — a performance that exceeded Luo Yonghao's own previous live appearances. The event was also reported as generating approximately $7.65 million in revenue for a Chinese entrepreneur operating through this model. Following the success of this deployment, more than ten celebrities and influencers began commissioning their own digital replicas for comparable commercial use. Baidu's e-commerce platform and Huiboxing (慧播星), a digital human collaboration platform, both reported a fourfold surge in inquiries in the period following the event. Douyin (抖音) subsequently introduced platform-level restrictions prohibiting the use of digital human hosts that do not engage interactively with viewers, marking a policy shift toward requiring active virtual participation. Broader coverage across NetEase and Sina documented the wider industry pattern, noting that AI-generated virtual hosts offer operators cost advantages alongside high customizability in voice, appearance, and behavioral style across diverse media formats.

Public Services: The Yunnan Medical Insurance platform has integrated digital human technology in combination with large language models and AI voice systems into its public service portal. Virtual agents deployed on the platform assist users through intelligent customer support, guided navigation, and video-based help services, constituting a concrete application of digital human infrastructure within government-administered healthcare-adjacent service delivery.

Cultural Production: Beijing Television and Hangzhou-based Xiaoxin Technology jointly created two 3D child-like avatars with rounded proportions for public exhibition at the World Garden Expo. These avatars incorporate lip-syncing, realistic facial expressions, and emotional engagement capabilities, demonstrating growing technical sophistication in the design of virtual beings intended for real-time, public-facing cultural and civic environments.

Market Overview: According to IDC's 2024 assessment of China's AI digital human sector, the market reached a scale of 4.12 billion yuan with year-over-year growth of 85.3%. Baidu Intelligent Cloud (百度智能云) was identified as the leading vendor across both infrastructure capability and deployment versatility. The assessment characterized digital humans as increasingly vital connectors between virtual and physical domains, with technical architectures undergoing broad upgrades driven by generative AI and large-scale model integration.

Research: Scholarly investigation into China's virtual influencer and AI anchor ecosystems has produced several relevant empirical findings. Y. Hongtao and Z. Guo compiled a corpus of over 35,000 posts by virtual influencers to examine how interactive response behavior affects user engagement, finding that active engagement by virtual influencers promotes measurably greater audience participation. X. Yan and co-authors examined the commercial impact of AI anchors and virtual influencers in live streaming e-commerce, emphasizing how AI-driven virtual personalities improve operational performance on Chinese platforms. Car, Jakopović, and Nehring specifically identified China as a leading adopter of virtual influencers for advertising and digital communication purposes, noting that the messaging of these entities is shaped entirely by the values and objectives embedded by their developers, a factor they flag as raising regulatory and ethical concerns in the context of digital information environments.

June 20 News

Livestream Commerce: Baidu (百度) deployed digital humans at scale within China's live-stream retail ecosystem, releasing Digital Human 3.0 in April 2025 as a product milestone for intelligent e-commerce. The centerpiece deployment was a digital human modeled on internet entrepreneur Luo Yonghao, introduced as the industry's first AI super-host, which outperformed its real-world counterpart in sales within 26 minutes of debut and generated over 55 million RMB in GMV across a seven-hour session, with Baidu presenting four major technical advances underlying the avatar's persuasiveness at its AI Day event. During JD.com's (京东) 618 shopping festival, over 17,000 brand merchants deployed Baidu digital humans alongside 14,000 concurrently operating AI agents, with total viewership exceeding 50 million. A separate Baidu AI avatar livestream earned USD 7.65 million in sales across more than 13 million viewers. Industry data cited in associated coverage indicates that digital human hosts in China reduce production costs by up to 80% and increase sales by 62% relative to conventional formats. Empirical research by Tang, Song, and Chung (2025), grounded in data from China and applying presence theory alongside the heuristic-systematic model, further demonstrates that AI-powered virtual streamers in tourism live commerce significantly shape viewer attitudes and increase continuous watch intention, reinforcing the strategic value of virtual beings in hospitality-oriented e-commerce.

Entertainment: The AI artist TaTa has been introduced as China's first self-described "pure-blooded" AI performer, distinguished from earlier virtual idols and 3D-modeled characters by being developed entirely through AI technologies for autonomous artistic expression rather than modeled on a pre-existing human or animation template. In animated film production, "A Better Tomorrow: Cyber Border" has been positioned as the world's first AI-powered kung fu animated feature, marking the integration of synthetic characters into Chinese narrative media at an industrial production level.

Cultural and Narrative Applications: Chinese companies are embedding digital humans within gaming and metaverse platforms to carry culturally specific narratives for younger audiences. Traditional heritage elements and intangible cultural skills are integrated with AI-powered virtual characters — including figures presented in Hanfu — deployed across gamified and immersive environments. These virtual beings function as cultural carriers, expressing Chinese identity and heritage through modern interactive interfaces that domestic commentary characterizes as new modes of narrative communication.

Marketing and Brand Representation: TikTok, a product of ByteDance (字节跳动), expanded its generative AI toolkit to allow brands to deploy highly humanlike digital avatars as synthetic marketing representatives across campaigns, positioning such avatars as a standard commercial instrument for audience engagement. Academic investigation into this commercial phenomenon includes Tang, Hao, and Li's 2025 study, which applies the Meaning Transfer Model to compare the effects of real versus virtual celebrities in livestream marketing contexts, finding that virtual celebrities influence viewer engagement and purchase intention in ways that differ measurably from human influencers.

Immersive Production and Technology: Technology companies in Guangdong are developing workflows that combine XR production pipelines with live-action shooting to achieve high-fidelity immersive visuals incorporating digital humans. These pipelines position digital humans not only as on-screen presences but as structural elements of enterprise-level multimedia production, requiring close integration between real-time rendering environments and live capture systems.

Ethical and Legal Contexts: A civil case in China established that using AI software to generate a virtual human from a natural person's likeness without consent constitutes a violation of personality rights, creating a legal precedent for image rights protection in the context of synthetic character creation and deployment. A separate report from Taiwan documents the psychological and legal risks associated with virtual emotional relationships and AI companion systems, specifically highlighting the exploitation of personal image in relation to youth populations exposed to synthetic personas.

June 19 News

Live Commerce: Baidu (百度) has deployed digital human technology at commercial scale in China's live commerce sector, with its most prominent case being a digital replica of entrepreneur Luo Yonghao. This digital human conducted a seven-hour uninterrupted livestream on Baidu's e-commerce platform, drawing over thirteen million viewers and generating more than fifty-five million yuan in gross merchandise volume. Baidu describes its digital human development as progressing through four distinct phases: early virtual appearances, hyperrealistic rendering, dynamic persuasive personality, and script-driven autonomous operation. The current iteration combines multi-modal interaction, emotional expression, and customized script generation powered by Baidu's Wenxin 4.5 Turbo large language model, representing the commercial displacement of human influencer roles by AI-generated synthetic characters in retail livestreaming.

Entertainment and Media Production: Shandong Digital Culture Group deployed its ultra-realistic digital human Qi Feng (齐风) as a co-host at a scene-matching event focused on embodied intelligence held in Jinan, where the virtual figure presented alongside a human host to launch project demonstrations and conduct thematic talks. This case illustrates the use of photorealistic digital humans in ceremonial and cultural hosting roles, including promotional presentations and international exchange contexts. Separately, director Zhang Wei produced the XR-enhanced media program Feng Qi Dong Fang (风起东方) featuring digital human character Li Li (厘里), completed over a two-month production cycle requiring precise coordination between extended reality assets and live-action filming to achieve a convincing three-dimensional effect. This production reflects the expanding integration of virtual human characters into Chinese television and film, including the technical demands of combining synthetic performers with physical sets in real-time. The Chinese mobile game Shadowverse: Extraordinary World (影之诗：超凡世界) offers players the ability to enter its virtual world through customized avatars, enabling social and competitive interaction with other players within an immersive narrative environment.

Education and Public Services: The AI digital human Mei Mei (梅·美) functioned as a virtual teacher and event host at a school activity in Wuxi, delivering the event agenda with fluent speech and engaging a live audience, as reported by Wuxi Daily (无锡日报). This deployment illustrates the adoption of AI-powered virtual humans in Chinese pedagogical environments as scalable facilitators capable of addressing large student groups. BytePlus (字节跳动), the enterprise technology division of ByteDance, released a demonstration showing the creation of a lifelike digital human voice designed to capture audience attention, convey complex emotions, and support brand communication across public services, education, and healthcare sectors in China. Together, these cases reflect a growing operational role for synthetic characters in Chinese public-facing institutional settings beyond entertainment or commerce.

Industry and Market Landscape: A market analysis published by Jfdaily (解放日报) estimated that China's virtual digital human sector carries a market potential exceeding six hundred billion yuan, attributing this growth trajectory to advances in artificial intelligence and virtual reality. The report identifies the emergence of large language models as a key driver of the intelligent evolution of digital humans, enabling broader adoption across retail, education, and enterprise communication. Virtual humans are characterized in this analysis as strategic assets for revitalizing consumption and augmenting workforce capacity, particularly in roles requiring scalable humanlike interaction at volume.

June 18 News

Marketing and E-Commerce: TikTok expanded its generative AI marketing suite, Symphony, to include dynamic digital avatars and AI influencers, enabling brands to upload images or short prompts to automatically generate branded video content featuring avatars capable of holding products, delivering scripted dialogue, and simulating influencer behavior. Baidu (百度) launched two complementary digital human initiatives targeting e-commerce live streaming: the Mengdie Plan, focused on producing high-profile "super-head" digital human streamers backed by traffic and promotional budget, and the Fanxing Plan, which aims to generate 100,000 free digital humans under the Huiboxing brand. The digital human of Luo Yonghao, a former technology entrepreneur, was deployed in a record-setting live commerce broadcast on Baidu's e-commerce platform, with a single stream drawing 13 million viewers, generating over 55 million RMB in gross merchandise value, and involving AI systems that processed 13,000 knowledge base items, produced 97,000 words of product content, and executed over 8,300 gestures through two digital human avatars. The Luo Yonghao avatar was subsequently integrated into Baidu's first dual-avatar interactive livestream room, enhanced by the Wenxin 4.5T model, where it demonstrated creativity, personality, and multi-modal interaction with high fidelity in replicating human expressions, speech, and behavior. Baidu also partnered with Pinscreen China (相芯科技), a face modeling company, to improve the performance of its AI streamers in e-commerce contexts, with the collaboration enabling AI anchors to match the performance metrics of human hosts; one beauty brand attributed 47% of its sales during off-peak hours to the adoption of these AI anchors.

Enterprise Training: Ning Shui Group (宁水集团) deployed the virtual digital human "表帅" on its proprietary intelligent training platform, Ning Shui Zhixing Academy, to deliver customized brand and cultural content tailored specifically to the water utility industry. The system leverages large-scale data models to support internal training, brand communication, and customer engagement, illustrating how sector-specific digital humans are being embedded into corporate learning infrastructure in China.

June 17 News

Live Commerce: Baidu (百度) deployed a digital human replicating internet personality Luo Yonghao on its e-commerce platform for a live commerce broadcast that attracted over 13 million viewers and generated more than 55 million RMB in gross merchandise volume, surpassing the performance of traditional livestream formats. The AI figure reproduced Luo Yonghao's voice, intonation, and facial expressions, and was paired with a second assistant digital human for synchronized co-hosting, marking an industry milestone in dual digital human broadcasting. A separate case, not attributed to a named company, involved two digital human hosts collaborating across a six-hour livestream that outperformed a human host's one-hour sales record within just 26 minutes. Baidu reported that more than 100,000 digital human hosts are now active across sectors including e-commerce, education, and healthcare, with the technology reducing livestream operating costs by over 80% and increasing gross merchandise volume by an average of 62%.

Marketing and Fashion: A Chinese AI startup is providing trend forecasting services to major brands including Nike and Urban Revivo through avatar-based simulations and predictive fashion analytics. This application positions digital human technology as an analytical and commercial instrument within the fashion industry, enabling brands to model trend scenarios through synthetic character simulations rather than conventional market research methods.

Content Creation: The Zilu Team developed an AI-generated virtual persona modeled on the character known as Landlady from Stephen Chow's film Kung Fu, building a recognizable content brand around an AI avatar derived from Chinese popular culture. This case illustrates how AI image generation tools are enabling the creation of virtual persona intellectual property tied to established cultural references, making IP development increasingly accessible to smaller production teams.

Virtual Companions: AI-generated companions in China are characterized as emotional substitutes within a growing consumer trend, with female users comprising 55% of the audience in this sector. These systems simulate emotional reciprocity and personalized connection through virtual dating scenarios, with some users spending thousands of yuan per month on emotionally interactive experiences. The appeal is attributed to emotional isolation, high-stress environments, and limited real-life confiding opportunities. Virtual hosts (虚拟主持人) and guided storytelling are employed to immerse users in synthetic romantic narratives, a pattern linked to what analysts describe as the she-economy.

Automotive Integration: ZEEKR (极氪) integrated a virtual AI avatar named Eva into its smart vehicle interface, first launching the character in 2021 with the assistant now in its eighth over-the-air software update. Eva operates across user lifestyle, entertainment, and automotive scenarios as a comprehensive voice-based virtual assistant embedded within the in-car experience, representing one of the more sustained deployments of a named virtual being within a Chinese consumer hardware product.

Cultural Events: The 40th GLOBAL GODDESS International Pageant, China Region, incorporated digital human technology into its competition structure, with contestants participating through both physical performances and digital avatar representations showcased on an official metaverse platform. The event used AI-generated virtual modeling and interaction-based scoring systems, establishing a hybrid model that combines physical and virtual presence within a structured cultural competition format.

Ethical Contexts: China Economic Net published commentary raising concern over the growing indistinguishability between real individuals and hyper-realistic digital human replicas, noting that viewers frequently cannot determine whether video content features a human or an AI-generated clone. The piece specifically flagged risks of portrait rights violations and potential fraud in cases where celebrity likenesses are reproduced without explicit disclosure or consent, framing these as regulatory challenges requiring urgent attention as digital human realism continues to advance.

Technology and Production: Light Chaser Animation, in collaboration with Qiwei Technology, offers a full-stack platform combining 3D visualization and AI for the creation and deployment of digital humans, with applications spanning intellectual property management and virtual character licensing. Extended reality integrations involving digital human performances are also being trialed in Macao, where precise environmental calibration is required to achieve realistic and immersive outputs in XR settings that incorporate live synthetic characters.

June 16 News

Livestream Commerce: Baidu (百度) deployed its high-persuasion digital human technology to support the commercial debut of a synthetic avatar representing the entrepreneur Luo Yonghao (罗永浩) on its e-commerce platform, where the digital persona surpassed the GMV of a real-person stream within 26 minutes of going live. Kuaishou (快手) hosted a digital human avatar of the influencer Xinba (辛巴) in a dedicated livestream that generated over 4 billion yuan in sales, while the influencer Viya (薇娅) separately launched an AI avatar in a fashion show-format broadcast. In Suzhou, Juchong Technology deployed a female digital human host during the "6.18" shopping festival, enabling continuous and seamless real-time interaction with customers throughout the commercial stream.

Enterprise and Public Services: iFLYTEK (科大讯飞) presented virtual digital humans as a core component of its product suite at a recent upgrade conference, integrating them alongside its AIUI platform and Robot Superbrain under a framework driven by large language models. The company deploys its digital human systems across customer service, education, and smart assistant applications, positioning them as a central interface layer within its broader enterprise and public-facing AI ecosystem.

Content Creation: The Zilu team (子鹿团队) produced AI-generated virtual characters for short-form comedic video content on social platforms, combining image synthesis and voice cloning to build distinct persona identities designed to reinforce brand recognition and enable scalable personal IP at production volume. Alibaba (阿里巴巴) open-sourced its MNN TaoAvatar project, a pipeline for high-fidelity 3D avatar generation and real-time interaction optimized for mobile devices, with stated applications in live streaming and virtual social environments and an explicit emphasis on maintaining rendering quality under device constraints.

Cultural Heritage: AI-driven digital human avatars have been applied to cultural heritage initiatives with Chinese involvement, including collaborative projects along the Silk Road connecting China and Central Asia, where synthetic virtual personas are used to preserve intangible cultural heritage and support cross-regional dialogue. These deployments frame digital human technology not solely as a commercial instrument but as a vehicle for cultural continuity and international knowledge exchange.

Ethical Contexts: Surveys conducted in China indicate that more than half of respondents cannot reliably distinguish AI digital humans from real individuals, a finding that has intensified scrutiny of disclaimer visibility and disclosure practices across the industry. Documented cases of misuse include celebrity impersonation, the generation of disinformation, and fraud, underscoring the risks that arise as digital human technology becomes more accessible and the regulatory and ethical frameworks governing its deployment remain underdeveloped.

June 15 News

Commerce and Livestreaming: Chinese entrepreneur Luo Yonghao announced the debut of a digital human version of himself, referred to as 数字人老罗, on Baidu's (百度) e-commerce platform for a livestream sales event. The deployment combines Luo's likeness with AI-driven interaction capabilities, positioning the digital human as an autonomous commercial agent capable of conducting retail product promotion and blurring the boundary between celebrity endorsement and synthetic salesmanship. The economic risks of this model are illustrated by a case from Jiangsu Province, in which an individual invested 80,000 yuan in establishing a virtual digital human (虚拟数字人) livestream operation, only to have his streaming account banned before recovering the investment. This incident underscores the financial and regulatory precarity facing entrepreneurs seeking to monetize synthetic characters through Chinese streaming platforms, where policy enforcement remains unpredictable.

Market Infrastructure: According to a 2024 IDC report on China's AI digital human market, Baidu, Huawei Cloud (华为云), and Xiaoice (小冰) were identified as leading developers in the space, with the report defining AI digital humans as virtual beings possessing human-like appearances, perceptual interaction capabilities, and AI-driven expressive functions. These platforms provide foundational infrastructure and modeling frameworks that support deployment of digital humans across media, enterprise services, and consumer applications. Alibaba (阿里巴巴) has separately open-sourced a 3D digital human application that combines high-fidelity avatar generation with real-time interactivity optimized for mobile devices, targeting use cases including virtual customer service, livestream hosting, and AR-based social interaction. The open-source release lowers barriers to synthetic character deployment in consumer environments and extends the range of developers who can integrate digital humans into everyday applications.

Healthcare: A Chinese research project developed a remote self-reporting platform that integrates multimodal digital biomarkers, including voice patterns and facial features, with direct patient input to assess and monitor mild cognitive impairment. The study frames these biomarker-driven systems within a "数字生物" (digital being) conceptual framework, indicating a trajectory toward human-like AI entities that interact with patients in diagnostic and monitoring roles. This approach signals an emerging overlap between virtual human interface design and medical applications in the context of neurodegenerative conditions, where synthetic interaction can augment clinical observation at scale.

Cultural and Educational Contexts: At a seminar hosted by Shanghai Xingjian Vocational College, virtual digital human (虚拟数字人) applications were demonstrated across cultural tourism and education, including AI-guided virtual explanations, QR-based navigation, smart guide maps, and metaverse exhibition halls. These deployments reflect how Chinese institutions are integrating synthetic characters into heritage presentation and public education, with an emphasis on service personalization and culturally situated storytelling. AI-driven avatar reconstructions of cultural sites in China have also been reported as tools for documenting and communicating cultural memory through immersive digital environments, extending the reach of virtual beings into heritage preservation.

AI Companions: Academic research published in Frontiers examined Chinese women's romantic relationships with AI companions, identifying emotional support and selective vulnerability as the primary motivations driving users to maintain sustained synthetic relationships. The study situates AI companions within a pattern of deliberate emotional engagement rather than passive use, pointing to the depth of attachment that virtual beings can generate in companionship contexts. This finding aligns with broader data indicating that a significant proportion of users across the Asia-Pacific region rely on AI companions for emotional advice, establishing synthetic characters as mainstream instruments for psychological support in China and neighboring markets.

Ethical and Legal Contexts: CCTV News (央视网) published a report raising ethical and legal concerns about the use of digital humans in China, citing the increasing difficulty for viewers to distinguish between real and synthetic personas and calling for stronger technical standards and regulatory oversight. The report highlighted documented instances of digital humans being deployed in fraudulent activities, demonstrating how synthetic characters can be weaponized in corporate and criminal settings when governance frameworks are absent. A report in 36氪 separately critiqued the practice of deploying synthetic characters with hyper-sexualized or boundary-pushing personas, noting that the companies producing such virtual identities retain full intellectual property over characters that have no real-world counterparts, raising unresolved questions about accountability, consent, and ethical governance in China's virtual being industry.

Media Production: An industry report on extended reality production referenced the character 厘里 as a case study in the complex integration of XR environments with digital human characters. The case describes how combining real-time XR materials with precise filming techniques is necessary to maintain visual coherence and technological immersion, illustrating a specific production challenge that arises when synthetic characters are embedded in spatially computed media. This highlights a growing class of virtual being deployment in Chinese media production where spatial computing infrastructure and synthetic character systems must operate in close coordination to produce commercially viable results.

June 14 News

Enterprise Applications: Alibaba (阿里巴巴) launched the open-source MNN TaoAvatar platform, a 3D digital human system built on a lightweight deep learning inference framework integrating 3D Gaussian reconstruction, AI-driven voice synthesis, and edge-device-compatible large models. TaoAvatar is deployable as a virtual customer service agent or virtual host, representing a commercially oriented digital human interface designed for scalable enterprise adoption. iFlytek (科大讯飞) expanded its digital human ecosystem by combining its AIUI, Robot Superbrain, and Xingchen platforms to develop AI virtual humans targeting enterprise-level smart interaction, with specific applications in office automation, human resources, and broader enterprise digital transformation workflows.

Market Landscape: IDC's June 2025 report on China's AI digital human market identifies Baidu (百度), Huawei Cloud (华为云), and Xiaoice (小冰) as key players in the sector. The report characterizes AI digital human systems as increasingly essential connectors between virtual and physical environments, reflecting the sector's maturation from a niche technology to a recognized industrial component within China's broader AI strategy.

Marketing and Live Events: Chinese entrepreneur Luo Yonghao deployed his digital avatar "Digital Luo" in a livestream event on June 15, with both the human and virtual versions appearing simultaneously, representing an emerging approach in commerce-driven influencer marketing in which digital humans extend the reach and productivity of real individuals. At Computex 2025, graphics card manufacturer Palit employed an AI virtual human named "Camila," developed in collaboration with Feiteng Image (沸騰映像), as its 37th-anniversary brand ambassador. Camila was designed for real-time interaction and on-site brand representation, illustrating the integration of synthetic characters into live commercial events for experiential branding purposes.

June 13 News

Livestream Commerce: The digital human application of prominent Chinese influencer Luo Yonghao (罗永浩) represents a high-profile convergence of top-tier IP and AI-driven retail. His digital human counterpart, built on Baidu's (百度) digital human technologies, was scheduled to debut on June 15 via Baidu E-Commerce, with preview sessions drawing over ten million viewers and generating GMV exceeding 50 million yuan in a single session. Luo framed the initiative around product authenticity, inviting viewers to serve as "human lie detectors" for the digital version, positioning user trust as central to the format's commercial viability. The wider deployment of digital hosts in Chinese livestream commerce has also encountered technical and regulatory vulnerabilities, illustrated by the "meow meow" incident, in which a digital host was manipulated through text commands to exhibit erratic on-stream behavior, exposing gaps in both legal safeguards and technical controls for AI characters used in commercial broadcasting.

Public Services: In Heilongjiang Province, the Pingfang District government deployed a digital human named "Pingping" at its administrative service center to handle citizen interactions and administrative inquiries in a physical public venue. This implementation reflects a pattern of municipal-level adoption of virtual beings in China, with local government entities turning to AI-driven digital human agents as a means of improving service efficiency and public accessibility.

Finance: Investor sentiment in China's capital markets is being shaped by corporate positioning around virtual digital human and AIGC concepts. Both Hengyin Technology and Chuanwang Media recorded notable stock surges linked to their stated affiliations with virtual digital human development and AI-generated content applications. These market movements indicate that virtual human technology is functioning not only as a discrete product category but also as an innovation signal capable of influencing equity valuations.

Legal Contexts: A landmark legal case in China addressed the status of a 中之人—the real person who operates behind a virtual character—and established a precedent for calculating damages based on the continuity of identity between an avatar and its human operator. The ruling emerged from a service contract dispute and evaluated value loss in terms of the virtual identity's sustained integrity, representing an early instance of judicial recognition of the economic and legal standing of virtual personas within China's legal system.

Enterprise and Technology: ByteDance's (字节跳动) PICO XR platform has been recognized for enterprise-focused extended reality capabilities, including full-avatar embodiment and remote training use cases in professional settings. At the developer tooling level, DeepSeek-Unity, an unofficial Unity package integrating DeepSeek's language models, enables the creation of intelligent digital humans within Unity projects, supporting text and voice interaction and lip-syncing features for non-player characters and interactive virtual agents, extending DeepSeek's capabilities into real-time character applications.

Cultural: Labubu, a virtual creature originating from the "Monster Forest" IP, has achieved broad international commercial traction through what analysts describe as a "borderless aesthetic" that departs from culturally specific East Asian visual conventions, contributing to its success across global markets and positioning it as an example of how virtual beings from China's IP ecosystem can operate as transnational cultural icons. Broader analysis of virtual streamer development in China has also identified the uncanny valley effect as a persistent technical challenge, with slight mismatches in facial expression or vocal intonation generating viewer discomfort and underscoring the continued need for refinement in avatar presentation to meet audience perceptual expectations.

June 12 News

E-Commerce and Livestreaming: Baidu (百度) has deployed its "慧播星" digital human livestreaming system as a comprehensive e-commerce solution for Macau-based retailers seeking to reach mainland Chinese consumers while managing high operational costs and cross-border regulatory differences. The system uses AI-driven digital human presenters to minimize reliance on physical production space, personnel, and equipment, and reportedly contributed to a 281% increase in livestreaming hosts and a 177% rise in monthly buyers over the course of a single year. Digital human adoption for livestreaming has also reached individual vendors: market vendor Fu Jiangyan in Yiwu has used AI digital human tools to drive customer traffic without requiring personal on-camera appearances, reflecting the scalability of such tools down to the smallest commercial operators. Academic research situates these deployments within a rapidly growing field. Chen and Yang have categorized virtual streamers into task-oriented and identity-based types, examining how this distinction shapes consumer purchase intentions according to product category. He and Zhan apply the appraisal-emotion-action model to analyze how AI streamers affect rural e-commerce audiences in China, linking perceived streamer attributes to downstream consumer behavior. Chen et al. compare AI and human streamers directly, finding that AI-driven presentations influence purchase intention through distinct cognitive and emotional pathways. A bibliometric study by Gong and Liu documents rapid growth in virtual streamer-focused research, with China accounting for a disproportionate share of output, reflecting concentrated academic and commercial attention on livestreaming as a primary deployment context for virtual beings.

Entertainment and XR Performance: Virtual hosts and augmented and extended reality-enhanced performances have become established components of live entertainment production in China. XR theatre productions including adaptations of "Macbeth" and "嵇康" incorporate real-time overlays of virtual elements such as dragons and fantastical set pieces, rendered through multi-camera systems to achieve immersive audience experiences. Festival events similarly employ this format, blending live performance with digitally generated characters and environments. These productions depend on meticulous on-site calibration and XR integration to sustain technological realism across live storytelling contexts.

Governance and Public Services: Several Chinese municipal governments have integrated digital human technology into public communication and administrative service delivery. Shenyang's Heping District launched "和智宣," a generative AI-powered digital human spokesperson developed under the district's "智慧和平" governance initiative to support government publicity and citizen-facing services. In the financial services sector, Synapse Technology (突触科技), a company operating under Derlin Holdings and having recently secured Series A funding from GPTX, C Capital, and Shuren Education, is developing AI-powered virtual customer service agents designed for deployment in dynamic financial communities.

Patents and Industrial Infrastructure: Beijing Jiuling Times Film and Media Co., Ltd. has received a patent for a "Metaverse Virtual Digital Human Production System," marking concrete progress in the industrialization of tooling for virtual human creation. Shenzhen Phantom Future Information Technology Co., Ltd. has separately secured a patent for a "virtual digital human interaction device, system, and method," advancing the technological basis for interactive virtual agents. Together, these filings reflect China's sustained investment in intellectual property infrastructure supporting the production, customization, and deployment of synthetic characters across multiple sectors.

Marketing and Consumer Reception: An AI-generated advertisement produced for Nongfu Spring's brand Oriental Leaves drew widespread negative public reaction due to the digital spokesperson's unrealistic visual rendering, most notably distorted finger representations. Viewer responses illustrated the uncanny valley effect in commercial digital human deployment, where insufficient visual fidelity undermines both viewer acceptance and brand perception. The case has become a reference point in Chinese industry discourse on the aesthetic challenges facing digital human producers and the production standards required for effective consumer-facing applications.

Research: Several academic studies examine the cultural, semiotic, and behavioral dimensions of virtual beings in Chinese contexts. Zhong, Li, and Morrison apply a triadic symbol model to analyze how digital humans in the metaverse function as vehicles of cultural communication, with particular attention to tourism and hospitality applications in China. Liu, Lin, and Liu assess how virtual influencers shape tourism purchase intentions differently from human influencers, contributing to the understanding of synthetic characters in destination marketing. Qiu et al. examine how anthropomorphic design in virtual idols affects emotional resonance and parasocial engagement, with findings applicable to both entertainment and marketing contexts. Gong and Liu additionally analyze virtual streamers as a primary application domain within AI-generated content research, documenting the expansion of this segment within the Chinese academic and commercial landscape.

June 11 News

Entertainment and Media: AI-generated digital humans have achieved significant visibility on Chinese short video platforms, with the figure Wang Jianguo (王建国) serving as a prominent case. Wang Jianguo is a synthetic character deployed to emotionally engage audiences, particularly older female users, and to stimulate consumer purchasing behavior, with content producers leveraging low-cost, high-efficiency tools to generate such avatars at scale. On Bilibili (哔哩哔哩), an AI digital human singer trended on the platform's hot charts, reflecting growing public acceptance of synthetic performers in mainstream entertainment. AI digital humans are additionally being deployed as virtual anchors and 24-hour livestream hosts, operating continuously without the logistical requirements of real-world presenters, effectively replacing traditional livestream talent and dissolving conventional boundaries between AI-generated and human-led media production.

Enterprise: NetEase (网易) has highlighted enterprise adoption of AI digital humans under the framing of "super employees," offering uninterrupted service while reducing costs associated with live staff. Detailed deployment guides for building digital human interaction systems have emerged to lower the technical barrier for startups and small and medium-sized enterprises, indicating that this technology is no longer limited to large-scale operators. These developments position AI digital humans as scalable front-facing agents embedded in customer service and retail engagement, substantively transforming front-office business operations.

Technology: Graphen AI (图策科技) presented its Aiia digital human series at COMPUTEX TAIPEI 2025, demonstrating AI-powered interactive solutions designed for smart engagement applications. The showcase attracted significant international attention and signals Taiwan's expanding footprint in the development and commercialization of virtual being technology.

XR and Tourism: The integration of digital humans into extended reality environments involves coordinated technical workflows encompassing scene shooting and digital material synthesis to preserve realism across production pipelines. The digital human known as "厘里" has been developed specifically for deployment in XR content creation contexts requiring this level of technical coordination. A fictional digital human named "木古木古" has been employed in tourism narratives, appearing across multiple locations as a place-based storytelling device embedded in tourism marketing content.

Education and Policy: Chinese entrepreneur Liu Dejian proposed at a UNESCO session the application of AI, augmented reality, and avatar technology to expand access to global education, framing synthetic agents as scalable instructional tools suited particularly to remote and under-resourced regions. This intervention represents a policy-level endorsement of virtual beings in educational contexts at an international institutional forum.

Ethical Contexts: The deployment of AI-generated digital humans in emotionally manipulative contexts has drawn criticism, with the Wang Jianguo case foregrounding the risks of using synthetic characters to exploit consumer trust among vulnerable demographic groups. Broader debate has emerged in China over the continued necessity of "中之人" — real human performers operating behind avatars — as AI-generated characters grow increasingly capable of replicating human expressiveness without any human contribution. These concerns are compounded by deployment scenarios in which audiences have no reliable means of distinguishing synthetic from human identities, raising unresolved questions about deception, exploitation, and authorship in AI-generated media.

June 10 News

Tourism and Cultural Heritage: China Mobile's (中国移动) Ganyu division, operating in Lianyungang, deployed a virtual digital human as a smart tourism assistant built using speech synthesis, emotion computing engines, and 3D holographic projection to interact directly with visitors and enhance travel experiences at the destination. A separate production titled "Our AI Virtual Human Goes to Suzhou," released on YouTube, showcased an advanced AI virtual human navigating and engaging with cultural sites in Suzhou, demonstrating the application of digital humans as media-facing tourism narrators embedded in real geographic and heritage contexts. At a forum on intangible cultural heritage, Donghua University (东华大学) and Lingdi Technology (灵迪科技) jointly presented a mixed-reality fashion show featuring digital humans alongside live models to reinterpret traditional Chinese textile arts — including brocade, embroidery, and tie-dye — positioning synthetic characters as culturally grounded interpreters of historical aesthetic heritage and bridging artisanal tradition with immersive technology.

Entertainment and Live Streaming: The live streaming sector in China has seen substantial expansion in digital human participation, with reduced technological barriers enabling a wider range of operators to deploy interactive digital human anchors across commerce and entertainment. Several large Chinese technology companies have entered this space and diversified application scenarios accordingly. A widely reported incident involved an AI digital human anchor operating under the "潮流夜生活" (Trendy Nightlife) livestream channel being manipulated through jailbreak prompts into performing as a catgirl and meowing repeatedly, exposing the tension between entertainment-driven customization and the need for behavioral controls over virtual personalities in participatory broadcast environments.

Marketing and Commerce: 引力传媒 (Yinli Media) registered a 0.9% sector gain on June 9 within the virtual digital human stock category, reflecting investor recognition of the commercial value attributed to synthetic character-based advertising and brand engagement. Virtual beings in this context are treated as monetizable assets capable of delivering consistent, identity-driven messaging across campaigns. Research by P. Shukla et al. in "Luxury: Where Next?" identified a strategy among luxury brands of pairing virtual influencers with culturally resonant contemporary narratives specifically to engage middle-income Chinese consumers, leveraging digital persona management to convey aspirational value and cultural relatability as a deliberate market positioning technique.

Education: The Chinese children's learning application 叫叫 (Jiaojiao) underwent a major upgrade introducing customizable personal virtual avatars that can form social "enthusiasm squads" to motivate learning through peer-oriented interaction. The integration of virtual personas into the application's pedagogical framework illustrates the use of emotionally resonant synthetic characters to drive engagement in early childhood education, combining gamification principles with avatar-based social dynamics to sustain learner motivation across sessions.

Technology: iFLYTEK (科大讯飞) filed a patent for a method of generating virtual character images specifically targeting the reduction of unnatural edge seams, a technical refinement designed to improve visual realism and surface continuity in rendered virtual beings. This development reflects sustained investment by a major Chinese AI company in the foundational visual fidelity infrastructure that underpins commercial digital human applications across industries including customer service, education, entertainment, and healthcare.

June 09 News

Cultural Heritage: SenseTime (商汤科技) collaborated with Sanmenxia Cultural Tourism Group (三门峡文化旅游集团) to produce what was described as China's first "Laozi Digital Human," a synthetic embodiment of the ancient philosopher Laozi rendered using computer vision, natural language processing, and 3D rendering technologies. The digital human is intended to function as an interactive historical guide within cultural tourism contexts, enabling visitors to engage with a lifelike AI-powered representation of a foundational figure in Chinese philosophical heritage. This project situates the digital human as a vehicle for national cultural storytelling and heritage preservation, combining advanced AI capabilities with the specific institutional goals of regional cultural tourism development.

Enterprise: Twinsverse Technology Co., Ltd. (孪生宇宙科技有限公司), a startup based in Hefei, has developed virtual digital humans for deployment across financial, legal, educational, and fire safety sectors. Co-founder Wang Zimiao, an alumnus of Anhui University, demonstrated these applications in the company's exhibition hall, illustrating the practical orientation of the platform toward public service infrastructure rather than entertainment. The breadth of sectors addressed reflects a deliberate strategy of embedding AI-powered virtual human interfaces into essential service roles that require consistent, accessible, and scalable human-like interaction.

Content Creation: Sina Finance published a ranking of top photo editing tools for 2024 that included a platform specifically serving VTuber creators in the Chinese content ecosystem. The featured tool supports advanced avatar formats including Hyper Avatars, VRM files, and Live2D models, and is designed to accommodate creators at both novice and professional levels of virtual persona production. The inclusion of this tool in a mainstream Chinese financial media outlet's ranking reflects the maturation and commercial legitimacy of avatar-based content creation as a recognized segment of China's digital media industry.

June 08 News

Entertainment: The virtual digital human "Ling'er" (灵儿), deployed on YY Live, operates through continuous 24-hour live broadcasting, dynamically responding to audience interactions by automatically triggering gift animations, adjusting background music, and modulating broadcasting rhythm based on real-time viewer sentiment analysis. By the end of 2024, Ling'er's introduction had measurably improved user engagement and retention metrics on the platform, illustrating the capacity of AI-driven virtual beings to sustain and deepen audience interaction in entertainment streaming contexts.

Enterprise: Shenzhen enterprises operating in the service sector have deployed AI digital employees across government and commercial operations, with reported productivity metrics indicating that a single digital employee performs work equivalent to that of ten human workers, yielding significant cost reductions. Shanghai Baihao Electronics has filed patents for a "Digital Virtual Human Intelligent Interactive Exhibition System," a technology designed to enable immersive user interaction within virtual exhibition environments.

Education: Minth University of Science and Technology (敏實科大), in collaboration with Giga Line (吉格線), launched Taiwan's first "Virtual Human Intelligent Interactive Webpage," an AI-driven platform designed to facilitate interactive engagements in educational settings as part of the institution's positioning as an "AI Professional University." The educational application Jiaojiao (叫叫APP) employs customizable virtual avatars to motivate learning, enabling users to personalize their avatars and form interactive "enthusiasm teams" that structure social learning environments and strengthen educational engagement through virtual identity representation.

Technology Development: Beijing Ultrapower Software (神州泰岳) has submitted patent applications describing methods for generating realistic digital humans using multimodal large model frameworks, advancing the technical foundations of AI-driven character synthesis. NetEase (网易) has been actively recruiting specialized AI engineers to develop and refine dynamic digital avatar systems, reflecting sustained corporate investment in digital human capabilities within China's technology industry.

Public Safety: Wuxi's immersive Ping An Experience Center has integrated AI-driven virtual avatars into security education programming, centering its approach on the "Truth Face" system, which is designed to expose fraudulent virtual identities within seconds. The deployment serves a public awareness function, familiarizing visitors with the mechanisms of digital deception and identity fraud through direct experiential engagement with virtual being technologies.

Ethical Contexts: The Cyberspace Administration of China has directed regulatory attention toward virtual beings and synthetic avatars, underscoring the need for compliance frameworks that ensure trustworthy and transparent AI deployments in this domain. These regulatory developments reflect an expanding institutional focus in China on algorithmic accountability, biometric privacy, and the governance of AI-generated characters, situating virtual being technologies within an increasingly formalized oversight environment.

June 07 News

Retail: Alibaba's Taobao Vision is deploying sophisticated 3D digital humans in flagship stores, enhancing customer interactions and enabling personalization within the e-commerce environment. JD.com has introduced digital human-led livestreaming capabilities that repurpose high-quality livestream content into targeted marketing snippets, amplifying audience reach and engagement across its platform. Hanhou Technology developed Feng Xiaozhi, an AI-powered digital human that facilitates intelligent and precise digital services for business parks, demonstrating the integration of virtual beings into enterprise-level customer service and marketing operations.

Cultural Preservation: Anqing Normal University has deployed digital humans to protect and disseminate Huangmei opera through a virtual reality museum system that incorporates immersive and interactive components, enabling users to engage directly with traditional performances in a digital environment. In Nanning, the AI virtual character Xiao Nan presents local historical archives through an engaging video series, revitalizing and popularizing regional history among the public by combining AI-driven virtual being technology with archival content.

Interactive Applications: Taiwan-based Minth University of Science and Technology partnered with AI startup Gigaline to launch an interactive webpage featuring virtual humans that integrate speech recognition, large language models, and virtual being technology, providing users with instant and dynamic communication capabilities as a demonstration of real-time generative AI applied to user-facing virtual being interfaces.

June 06 News

Entertainment and Gaming: Tencent (腾讯) is developing a full-stack AI game creation solution that incorporates interactive virtual beings capable of responding via text, voice, and video, indicating a strategic move toward generative and real-time AI characters at the platform level. NetEase's (网易) Fuxi AI Lab has introduced virtual beings and smart facial modeling technologies into its titles, including the title 倩女幽魂, applying digital human techniques to character rendering within established game franchises. At the Fusion Game Carnival 2025 in Beijing, organizers unveiled a range of artistically stylized virtual figures (虚拟形象) designed for narrative immersion and brand engagement, reflecting the sustained use of custom virtual characters as storytelling and promotional instruments within China's fandom and creative industries. Tencent's Hunyuan video avatar system, referenced in connection with the WanGP 5.4 framework, enables voice- and song-driven avatar generation of approximately fifteen seconds in duration using consumer-grade hardware, representing a significant reduction in the computational resources required to produce expressive digital characters in real time.

E-Commerce and Marketing: JD.com (京东) deployed advanced digital humans during its "618" shopping festival for livestreaming, product showcasing, and customer engagement, with implementations including custom avatars developed for the brand Three Squirrels (三只松鼠) and so-called "Digital Human 3.0" models equipped with dynamic emotional expression and real-time video slicing capabilities. Reports sourced from Aumedia (澳洲亿忆网) indicate that vendors in China can now generate realistic digital humans for commercial use at near-zero cost by training models on large datasets of real human images and applying voice-driven animation, with businesses deploying these avatars as digital employees across marketing, livestreaming, and customer support functions, with customizable facial features and vocal traits increasing their commercial versatility. The China Internet Association has forecast that the core market for digital humans in China will exceed 48 billion RMB by 2025, with marketing and customer service identified as primary application domains driving this growth.

Film and Media Production: Chinese filmmakers are now using AI-generated digital humans to replace real actors for entire sequences, a development reported in 科技日报 (Science and Technology Daily) that signals a structural shift toward cost-efficient, digitally mediated production workflows in China's film industry. This practice consolidates AI synthesis, facial rendering, and performance replication into a single production step, reducing dependence on physical cast presence and enabling studios to generate or modify on-screen characters at scale.

Cultural and Educational: Shanghai Lida College (上海立达学院) presented a virtual digital human representation of Dong Qichang, a Ming Dynasty painter and calligrapher, at the Hong Kong Cultural Expo as part of an educational project combining art history and AI-driven interaction. The project exemplifies the application of digital human technology to cultural heritage preservation and pedagogy, using a synthetic character to embody historical knowledge in an interactive format accessible to contemporary audiences.

Biomedical Research: The Guangzhou Institutes of Biomedicine and Health (广州生物医药与健康研究院) is applying the concept of digital lifeforms (数字生物) to describe AI-generated biological simulations including digital cells and biomarkers. While positioned within biomedical research rather than consumer-facing digital human deployment, this work extends virtual human frameworks into predictive health modeling, treating synthetic biological representations as analytical agents within clinical and research contexts.

June 05 News

Education: AI digital humans and virtual characters have been deployed across multiple educational contexts in China and Taiwan. In rural elementary schools in Inner Mongolia, AI digital humans enhance classroom environments and support extracurricular activities through interactive digital instruction. At Daqing Middle School, AI-based digital characters (数字人物) enable "dialogues across time and space," increasing student engagement in physics lessons. Research on VTubers at National Taichung University of Education in Taiwan examined the role of Kaku Kamu, a VTuber employed by the university, in fostering student participation during campus events; a survey of 300 students found that VTuber integration significantly enriched campus culture by bridging digital and physical interaction.

Enterprise: At the 2023 semiannual performance briefing of Xiamen Tungsten Industry (厦门钨业), a virtual host (虚拟主持人) powered by AI digital human technology presented financial data interactively, with the stated aim of making corporate disclosures more accessible to investors. This deployment illustrates the growing adoption of AI-driven virtual presenters in formal corporate communications in China.

Marketing: ByteDance (字节跳动), through its advertising platform Juliang Engine (巨量引擎), has launched a product called Jichuang (即创) that enables brands to generate AI-produced videos featuring digital humans, reportedly increasing content production efficiency while reducing costs. During China's 618 shopping festival, brands deployed "digital employees" (数字员工) at zero cost as front-line customer engagement tools, demonstrating the scale at which virtual beings are being integrated into live commerce operations. Companies including Zhongyin Fashion and Yuan Qidian have invested in core technologies such as 3D digital reconstruction and cross-modal AI interaction, reflecting competitive positioning in AIGC-plus-3D virtual being development; the virtual digital human market in China is projected to exceed 48 billion RMB by 2025. Academic research on AIGC-enabled cross-border e-commerce live streaming examines how AI-generated virtual streamers can operate continuously, adapt speech and visual behavior dynamically based on real-time feedback, and personalize content to viewer preferences, underscoring the commercial utility of synthetic characters in international e-commerce contexts.

Legal and Intellectual Property: Bilibili (哔哩哔哩) received a patent for a system enabling live streaming based on virtual avatars (虚拟形象), marking a step toward the formalization of technical and legal infrastructure around digital personalities in China. A separate legal case concerning a virtual streamer (虚拟主播) raised questions about whether avatar movements and voices qualify for intellectual property protection, and how the identity of the human performer behind the avatar—referred to as 中之人—relates legally to the virtual persona; courts and legal professionals are now addressing the implications of avatar-based contractual roles and the identity overlap between performer and character. The deployment of virtual beings in metaverse environments has further complicated trademark protection frameworks, as legal definitions are being extended to account for digital-only entities such as virtual digital humans (虚拟数字人).

Cultural: At the 2024 Hong Kong International Cultural and Creative Expo, Shanghai Lida University presented a digital human incarnation of Ming Dynasty artist Dong Qichang, allowing visitors to interact with the virtual figure via smart devices as a means of promoting traditional Chinese culture through AI visualization. This application is part of a broader pattern in which digital transformation performances—including holographic dance shows featuring virtual figures—are employed in China to present heritage content through immersive, technology-mediated formats.

June 04 News

Marketing: Jichuang (即创), ByteDance's AI content creation platform embedded within the Douyin commercial ecosystem, has been iterating on its digital human marketing product for over a year, applying video and voice model technologies to improve output quality, content diversity, and production efficiency at scale. Platform data indicate that users collectively generate more than 400,000 AIGC content pieces daily, with merchants relying on the digital human video completion capability to significantly shorten production cycles and reduce costs; forward-looking merchants have additionally unlocked the option to commission custom proprietary digital humans through the platform, effectively creating a dedicated virtual representative to work independently on high-volume tasks. In the live e-commerce context, JD.com (京东) released a live stream highlight clip function that uses AI to assist digital human anchors in repackaging high-performing live broadcast content into short-form video clips, enabling brands to maximize both traffic acquisition and downstream conversion from existing live sessions. Chanjing AI (蝉镜AI) operates a dedicated e-commerce digital human video creation platform for merchants who want to generate product-selling video content without on-camera human talent, providing a large library of e-commerce avatar types and automated shot composition tools to compress the production process. Jiuma AI Digital Human (九马AI数字人) takes a vertical infrastructure approach, offering high-precision digital human modeling alongside multi-scene virtual anchor services that support 24-hour unmanned live streaming and full live room construction; its AI-driven voice interaction and expression-and-motion simulation technologies are designed to produce near-human broadcast quality across e-commerce, tourism, and education deployments. The brand IP digital human solution provider jyfai.com offers a bundle combining CEO narration avatars, dynamic brand walls with festival-specific skin changes, and historical footage restored to 8K resolution; one automotive brand implementation using the platform received a design award and achieved a 22% improvement in private-domain conversion rates. DeepAI Digital Human Anchor (DeepAI数字人主播), developed by Hefei Jieliu Networking Technology (合肥市杰留资网络科技有限公司), is a dedicated application for virtual anchor use cases distributed through Tencent's Yingyongbao app marketplace.

Entertainment: Discussion within China's film and television sector about the commercial deployment of digital human avatars derived from named celebrities has become a visible point of industry analysis, with practitioners observing that such avatars have begun substituting for real actors in selected filming scenarios. Industry commentator Zhao Chunyan notes that as celebrity digital human representations take on partial shooting responsibilities, actor fees for those segments are expected to fall, with the resulting cost savings anticipated to be redirected toward film technology investment rather than talent remuneration. The broader commentary in Chinese media frames this development within a larger convergence of AI video generation and digital human production capabilities, with Kling (可灵), a video generation model developed by Kuaishou (快手), cited alongside digital human face-swapping and voice cloning tools as part of an integrated shift in how screen content is produced, edited, and circulated. Practitioners and observers characterize this combination of AI-driven animation, synthetic voice replication, and video manipulation as fundamentally restructuring the production economics of China's audiovisual content industry.

Technology: ByteDance (字节跳动) open-sourced LatentSync 1.5, its most capable AI digital human model as of 2025, designed to support one-click deployment and enabling users to generate lip-synced digital human video outputs in approximately five minutes. A community-enhanced distribution of LatentSync, documented and shared via Bilibili, incorporates a facial enhancement processing layer, extends hardware compatibility to graphics cards with as little as 6GB VRAM, and supports audio-driven video synthesis with full lip-form compositing for streaming inference workflows. Sonic is a separately described lightweight digital human lip-sync model that accepts a single portrait photograph and an audio input, generating a video clip with precise mouth-shape alignment and naturalistic facial expression suitable for virtual digital human production, virtual anchor content, and AI-narrated video commentary applications. An open-source project titled Ultralight Digital Human, published to GitHub by Chinese developer anliyuan, is characterized as the first open-source digital human model of its class capable of real-time inference on mobile devices, using a short recorded training video to produce personalized motion material for streaming deployment. Baidu Smart Cloud (百度智能云) provides a 3D digital employee product accessible through its open developer platform, targeting enterprises seeking to deploy interactive digital human staff in service or operational contexts. The AI Max mobile application, distributed on the Apple App Store for the China market, allows users to generate a personalized digital human agent from a self-recorded video clip and then produce talking-head content by inputting a script, enabling users who are unwilling or unable to appear on camera to produce continuous video output.

Enterprise: Sangfor Technologies (深信服) lists a sales digital human as one of its active commercial AI product lines, integrating it into a portfolio that also includes an AI computing platform, a domain-specific security large model, and embodied intelligence applications under an "AI First" strategic direction; its sales digital human has been applied in customer-facing sales workflows. China Communication University (中国传媒大学) organized a 2025 practical training workshop titled AIGC and Virtual Production Technology, covering digital human driving and XR virtual-real compositing as primary technical competencies alongside AI micro-course production and intelligent short-video matrix operations; the program projects that participants can achieve a 400% improvement in industrialized content production efficiency upon completing the training.

June 03 News

Marketing and Content Creation: ByteDance (字节跳动) operates Jichuang (即创), an AIGC-powered digital human platform embedded within the Douyin commercial ecosystem that has scaled to over 400,000 AI-generated content pieces produced by users daily. Merchants on the platform use digital human video generation tools to reduce production cycles and costs, while more advanced users commission bespoke private avatar personas to function as persistent branded representatives. Chanjing AI (蝉镜AI数字人) provides a dedicated e-commerce digital human video creation platform for online sellers lacking access to on-camera talent or production crews, offering category-specific templates including traditional Chinese wellness and cultural health themes. Keevx, an AI digital human video creation tool developed by a Beijing-based company for overseas small and medium enterprises and independent creators, combines AI script generation with multilingual translation and curated public avatar libraries to deliver one-click marketing video production for users without prior editing experience. Nami AI (纳米AI) version 3.0 supports single-person video generation in which users upload their own likeness and voice to produce e-commerce and self-media content across varied scenes, costumes, and movements. Creations Star (创作星), available on iOS, is an all-in-one AI content application centered on digital human video generation for spoken-word format videos and novel-to-video conversion, capable of producing a professional spoken-word video from a brief on-camera recording within approximately three minutes.

Live Commerce: Digital human anchor systems have emerged as a major deployment category within China's live commerce sector. BocaLive is a dedicated AI digital human livestreaming tool designed for simultaneous management of multiple platforms and storefronts, enabling merchants to run intelligent automated broadcasts that overcome the time and staffing constraints of conventional live shopping operations. Aihuoke (爱火客), which operates the aihuoke666 platform, provides a hyperrealistic 24-hour AI digital human livestreaming system marketing one-to-one real-person avatar cloning as its core capability, with the stated aim of driving precision traffic conversion through continuous unmanned broadcasts. Yizhi Intelligence (一知智能) has developed a digital human anchor system integrating a large language model described as an autonomous decision-making engine, enabling unmanned full-loop live commerce; the company claims the system achieves comparable real-person conversion rates at approximately two percent of conventional costs and has been identified as part of Hangzhou's agent innovation pilot initiative. JD Technology (京东科技), the technology subsidiary of JD.com (京东), has built its Yan (言) digital human and AIGC product suite around enterprise marketing use cases spanning digital human live commerce, AI livestreaming, and digital marketing, with deployments focused on intelligent customer service and marketing activation for enterprise clients.

Technology and Open-Source Development: Multiple major Chinese technology companies released or updated open-source digital human models and frameworks in June 2025. ByteDance open-sourced LatentSync 1.5, described as an audio-driven lip-synchronization digital human model supporting one-click deployment and rapid production; the model attracted substantial developer attention on Bilibili, where third-party contributors published enhanced distribution packages adding frame interpolation, face restoration, facial enhancement, and legacy GPU compatibility. Tencent (腾讯), through its Hunyuan team, open-sourced HunyuanVideo-Avatar, a multimodal generative model that produces high-definition video from a single uploaded image and audio input in approximately 14 seconds, with support for simultaneous multi-character rendering and emotion transfer targeting short-form video and e-commerce livestreaming applications. JD.com open-sourced JoyHallo, an AI digital human model designed specifically for Mandarin speech synthesis that embeds audio features via the wav2vec2 architecture and uses a semi-decoupled structure to improve lip-movement prediction accuracy, with additional support for English video output. Ant Group (蚂蚁集团) developed EchoMimicV2, a half-body digital human animation system generating video from a reference image, an audio track, and a hand gesture sequence to produce synchronized speech and natural upper-body movement. Silicon Intelligence (硅基智能) open-sourced both HeyGem, a short video generation tool capable of cloning a person's appearance from as little as one second of reference footage and completing output in 30 seconds, and DUIX, a real-time interactive digital human engine, with associated real-time rendering packages published under the duix-guiji namespace on npm. A GitHub project titled Ultralight-Digital-Human, released by a Chinese developer, claims to be the first open-source digital human model lightweight enough for real-time inference on mobile devices, using 20-second training video clips as the source material for streaming inference. OpenAvatarChat is an open-source modular interactive digital human framework supporting real-time conversation, multimodal understanding, and natural interaction on a single consumer-grade PC without cloud deployment or high-end server hardware. Sonic is a lightweight lip-sync model that generates video with precise mouth movement and natural facial expression from a single portrait photograph and an audio input, positioned for use in virtual digital human production, virtual anchor applications, and AI video narration.

Cultural and Tourism Applications: A digital human of the Taoist philosopher Laozi has been released as the first of its kind in China, deployed at a scenic heritage site in Henan province through an integrated smart terminal that enables visitors to conduct philosophical dialogue with the avatar character, exploring themes including the interdependence of fortune and misfortune drawn from the Tao Te Ching. Zhejiang Zhongnan Animation (浙江中南卡通股份有限公司) has developed a holographic AI digital human of the Song Dynasty poet and statesman Su Dongpo, now in its version 3.0 iteration; having undergone approximately one year of conversational training since an earlier less capable release, the digital human is described as substantially more fluent in dialogue, and the company has indicated plans to develop a further digital human of the historical figure Wang Anshi. In Jiangxi province, an AI-driven intelligent tour guide digital human named Xiaoxiaozi (小小子) has been deployed at a cultural heritage exhibition to lead visitors through displays, providing interactive explanations of historical artefacts including scroll paintings and carvings through natural conversational interaction, with the application framed in provincial press as evidence of the role digital human technology plays in the digital transformation of cultural tourism. The same Jiangxi cultural venue has also deployed metaverse technology enabling visitors to experience Tang Dynasty street environments and historical court settings through digitized avatar bodies, combining AI-character interaction with immersive heritage scene access.

Healthcare: The Shanghai Municipal Health Commission has issued a formal research call soliciting project proposals on the development of doctor digital humans and digital twin medical assistants, specifying use cases in condition monitoring and early warning, bedside intelligent nursing devices, and integrated patient service covering pre-consultation full-process management, cloud-assisted in-consultation care, and intelligent post-consultation follow-up. On the JD.com platform, an AI nutritionist deployed as a digital human character provides 7-by-24-hour personalized dietary consultation through multimodal interaction across text, voice, and video channels, generating customized phased nutrition plans based on individual health records and behavioral data, and is reported to have cumulatively served over ten million families. A technical analysis from Juejin describes a medical industry digital human agent architecture that integrates real human visual appearance, an AI reasoning system, and intelligent visual perception into a unified interactive entity, functioning as an intelligent assistant within clinical and health management settings under large-model AI technology, with the article framing this as a contribution to the convergence of industrial production and medical services.

Enterprise and Finance: Huawei Cloud (华为云) offers commercial digital human services that generate modeled avatar personas from user-provided video recordings, supporting virtual anchor, virtual host, virtual customer service, and virtual teacher configurations, with a voice customization layer to align synthetic speech characteristics to specified persona requirements. Sangfor (深信服), a Chinese cybersecurity and cloud infrastructure company pursuing an AI-first strategy, lists sales digital humans among its active AI product lines alongside large-model security tools and AI computing platforms, with the sales digital human application already in production deployment within enterprise client workflows. Zero-One Wanwu (零一万物), a Chinese AI startup, has restructured its business to concentrate on digital humans and B2B model customization and deployment after scaling back its consumer-facing product lines, with the digital human segment identified as the primary enterprise channel focus going forward. Beijing Shenghuan Julong Technology (北京盛华聚龙科技有限公司) has deployed its Zhifeng AI (智峰AI) digital human for a domestic private bank, replacing both human-staffed service and internet-model operations with AI-driven financial product interaction, and describes the institution as the only small-business financial services bank in China operating entirely through AI-agent service rather than conventional human agents or internet channels. A smart retail analysis from Eastmoney identifies 3D AI digital human virtual shopping guide screens as an immersive in-store experience layer within next-generation intelligent retail environments, described alongside AR fitting technology and electronically connected price-tag and mobile payment systems as components of AI-enabled physical retail.

Education and Training: Teddy Smart Tech (泰迪智能科技) operates an AIGC laboratory framework for university-level digital media programs that integrates digital human application development and intelligent agent design into a practical curriculum, aiming to cultivate professionals in new-generation digital media production through real-project training scenarios. Xiamen University's Database Laboratory has incorporated a formal experiment on using Tencent's Zhiying (智影) large model to generate digital human broadcast videos into its published textbook for digital literacy and AI general education, indicating that digital human generation workflows have entered structured instructional content at the Chinese university level. A government-organized AI enterprise empowerment training program run by the Shenzhen Municipal SME Service Bureau included dedicated sessions on digital human livestreaming and short video production, delivered by a JD Technology AIGC product specialist covering AI digital marketing and AI livestreaming SaaS tools, indicating that digital human commercial deployment has become a standard topic within government-organized enterprise AI adoption programs across China.

June 02 News

Entertainment: Tencent (腾讯) released HunyuanVideo-Avatar, a multimodal diffusion Transformer model co-developed by the Tencent Hunyuan team and Tencent Music Tianjin Lab that accepts a single image of a real person, cartoon, or 3D character along with an audio clip and generates an emotion-controllable, motion-rich short video at up to 720p resolution in approximately 14 seconds. The model supports multi-character on-screen dialogue and is fully open-source. Separately, a Chinese content creator workflow documented on Zhihu describes producing original songs using AI tools and then rendering them through AI female digital human avatars for distribution as original video content across major Chinese platforms, with some individual videos reaching over ten million views.

Marketing: Digital human livestreaming has become a structured commercial service in China, with published systems framing the offering not as simple video looping but as a complete digital human broadcast system capable of 24-hour uninterrupted automated content delivery, single-script deployment across multiple accounts, and real-time interactive support. Daiyun Zhisuan Livestreaming (戴云智算直播) positions itself as a full-domain solution deployable across Douyin, Kuaishou, Meituan, and Xiaohongshu for merchant self-use or individual livestream commerce. JD Technology (京东科技) has developed digital human livestreaming and short video SaaS products targeting enterprise marketing and intelligent customer engagement, with its AI digital human and AIGC product division presenting at a Shenzhen government AI innovation training session. Naomi AI (纳米AI) version 3.0, distributed via the Microsoft Store, offers a single-person video generation solution that uses an uploaded likeness and voice to produce multi-scene, multi-outfit broadcast videos for e-commerce and self-media; the product also supports one-sentence generation of micro-films and news readings with multiple characters and voices. The Chuangzuoxing (创作星) mobile app targets self-media creators with digital human lip-sync broadcasting, enabling a professional-quality talking-head video in under three minutes without any filming setup. A Beijing company launched Keevx, an AI-driven digital human video creation tool for overseas small and medium enterprises and individual creators, offering AI script generation, translation, curated portrait templates, and one-click marketing video production; the same product is also listed under the name Keeva AI on Chinese AI tool aggregators.

Platform and Tool Development: MoFa (Shanghai) Information Technology Co., Ltd. (魔珐（上海）信息科技有限公司) defines its 3D digital humans as distinct from traditional 3D characters in films and games such as Avatar and Black Myth: Wukong, and from virtual influencers such as Ayayi and Lil Miquela, by emphasizing real-time 1v1 interaction capability, live streaming support, and dynamic personalized content generation based on user input. Huawei Cloud (华为云) provides an end-to-end self-service cloud platform for digital human development covering model training through content generation, with batch training task management, visualization, and multi-channel concurrent inference using a single card. KreadoAI (可灵道AI) operates an online AI digital human actor platform offering template-based video production with a library of avatar samples. Chanjing (蝉镜) offers a Chinese-style digital human template named Wenhao (文昊) for knowledge-broadcasting short video production, available to subscribers for one-click video generation. The open-source GitHub repository awesome-digital-human-live2d provides a modular framework for building customized digital humans using large language models, with support for coding extension modules and agent orchestration frameworks. AIheygem, an integrated software package for digital human generation, released version 2.0 supporting batch processing and 50-series graphics cards without requiring Docker or WSL installation. A technical guide published in June 2025 documents a deployment stack combining Ollama, Docker, DeepSeek, and Dify to produce locally hosted voice-interactive digital humans, with performance scaling according to GPU capability. Framepack, a free open-source image-to-video tool, has been adopted within the Chinese digital human community as a supplement to HeyGem, which produces only mouth movement from a still image; Framepack enables more realistic full-body dynamic motion for the resulting digital human output. MorphMe (锐影), distributed via Chinese app stores, provides AI-driven character costume transformation including historical, retro Hong Kong, and other style transfers applied to human subjects in video.

Cultural and Tourism Contexts: Jiangxi province cultural tourism has deployed an AI-driven intelligent tour guide digital human named Xiao Xiao Zi (小小子) that uses interactive presentation to guide visitors through cultural artifacts including a twenty-four-merit painting and Song Dynasty floats. The same report describes metaverse applications allowing tourists to use digital avatars to move through reconstructed Tang Dynasty urban environments, markets, and court banquets, enabling intelligent interaction with virtual scenes and historical characters.

Healthcare: The Shanghai Municipal Health Commission issued a 2025 call for smart medical research projects that includes a dedicated track for doctor digital humans and digital twin assistants, specifying development of condition monitoring and early warning systems, bedside intelligent nursing devices, and AI-supported service functions spanning pre-consultation full workflow, cloud-accompanied consultation, and post-consultation intelligent follow-up.

Enterprise and Finance: Beijing Shenghuan Julong Technology Co., Ltd. (北京盛华聚龙科技有限公司) has deployed its Zhifeng AI (智峰AI) digital human product at a domestic private bank described as China's only small enterprise financial services bank operating entirely through AI intelligent service in place of traditional human and internet-based channels. Sangfor Technologies (深信服) lists sales digital humans as a current commercial AI application alongside its proprietary large model for security, AI computing platform, and embodied intelligence products, and has been actively recruiting AI engineers for digital human-related development roles.

June 01 News

Live Streaming: Chinese digital human technology has seen significant activity in live streaming applications, with multiple platforms and products enabling autonomous or semi-autonomous AI-hosted broadcasts. Zhituo Guangjie (智拓光界), operating at ztgj.com, offers an AI digital human live streaming system that supports online avatar cloning and claims to reduce live streaming costs by 90%. A digital human live streaming system described on the Juejin developer platform positions itself not as simple video playback but as a complete automated broadcast infrastructure, capable of 24-hour uninterrupted content output from a single script distributed across multiple accounts simultaneously. Zhibo AI (智泊AI), through its zhipoai.com corporate site, developed an AI virtual streamer product based on the open-source VirtualWife framework, framed as an intelligent, interactive virtual live broadcasting robot. Central Academy of Fine Arts AI Technology Xiong'an Co. (央美人工智能科技雄安有限公司) lists digital human live streaming software as one of its core products, with that system having received an excellence award at the Xiongcai Cup innovation and entrepreneurship competition. A Zhihu post describes an independent developer's effort to replicate the Microsoft Xiaoice (微软小冰) virtual human e-commerce live streaming system, noting that Xiaoice's commercial product requires manual customization of voice and appearance, has usage and duration limits, and is cost-prohibitive for individual operators, while the replicated version aims to deliver rapid self-service avatar and voice cloning without those constraints.

Content Creation: Several Chinese products focus on enabling individuals and small teams to generate digital human video content for short-form platforms without requiring filming equipment or on-camera presence. Chuangzuo Xing (创作星), available on the Apple App Store, is a multi-function AI creation app whose core feature is a talking-head digital human (口播数字人) module that uses an AI avatar to synchronize lip movement and generate a professional broadcast video within approximately three minutes from a spoken script. Nami AI (纳米AI), distributed via the Microsoft Store, offers its 3.0 version as a single-person video generation solution supporting e-commerce live selling and self-media talking-head content, using user-uploaded images and voices to produce videos across multiple scenarios, outfits, and dynamic actions. A Zhihu post documents a workflow in which AI tools are used to compose original songs and then AI digital humans perform and deliver those songs as original video content published across Chinese social platforms, with the creator reporting that individual videos in that format have reached over ten million views. Chongqing To Do Technology Co. (重庆多多科技有限公司) publishes a suite of related apps including Koubo AI (口播AI) and AI video creation tools targeting the same talking-head digital human content production use case.

Technology: Chinese technology companies have released and open-sourced foundation models specifically designed for digital human animation and avatar generation. Tencent (腾讯) released and open-sourced HunyuanVideo-Avatar, a multimodal digital human generation model jointly developed by the Tencent Hunyuan video large model team and Tencent Music's Tianqin Lab using MuseV technology; the model accepts a single character image—covering real people, cartoon figures, or 3D characters—plus an audio clip and produces video output at up to 720p resolution in approximately 14 seconds, with support for multi-character simultaneous rendering and emotion-controllable expression. Alibaba Ant Group (阿里蚂蚁集团) released EchoMimicV2, a half-body AI digital human animation system that takes a reference image, audio, and hand pose sequences as inputs to generate animated video, building on its predecessor EchoMimicV1. MoFa (Shanghai) Information Technology Co., Ltd. (魔珐（上海）信息科技有限公司) articulates its 3D digital human product category through its Youyan platform, distinguishing its characters from conventional 3D game or film avatars by their capacity for real-time video, live streaming, and one-to-one interactive dialogue, with the ability to generate personalized content in real time based on user requests.

Enterprise and Platform Deployment: The market for deployable digital human infrastructure targeting Chinese developers and businesses is active, with both commercial and open-source options. GenHuman, hosted on GitHub, is an API-based digital human product packaged for one-click commercial deployment, including a web interface, mobile application, WeChat Mini Program, and backend management console. A Bilibili tutorial demonstrates one-click cloud platform deployment of a real-time voice dialogue digital human combined with a live streaming version, integrating GSV V2 and Ollama, released as a customized package by Luoxi AI (洛曦AI) running on cloud GPU infrastructure. Baichen Group (百辰集团) markets Huawei Cloud (华为云) digital human services through its platform, offering an end-to-end self-service cloud interface for model training and content generation with support for multi-card concurrent inference. A Beijing-based major technology company identified only by the reporting outlet launched an overseas AI digital human video creation product named Keevx, targeting small and medium-sized enterprises and individual creators in international markets who lack video editing experience or production teams, with features including AI script generation from topic descriptions and digital human avatar-driven video output. ByteDance (字节跳动) launched Xiaoyunque AI (小云雀AI), a product capable of one-click digital human video generation and design image creation from text or image inputs.

Marketing: Digital human synthetic characters are being deployed in Chinese marketing and promotional contexts as a replacement for human spokespeople. An Instagram post circulating in Chinese-language digital marketing communities promotes the use of digital human avatars with fixed AI-generated identities to operate personal IP and content channels without on-camera appearance, described as a strategy for affiliate and e-commerce monetization. A Facebook post from a digital marketing account documents the use of AI digital human promotional videos as an alternative to conventional paid advertising, framing digital human content as a tool for improving commercial performance when traditional ad spending fails to meet targets. Yuan Xiang AI (元像AI), available on the Apple App Store, provides digital human-adjacent features including AI live streaming and video production for portrait-based content.