This article is written by Chiu Chaucer (Senior Deputy General Manager of Inventec / Deputy Chairman of the Technical Expert Committee of the Taiwan Cloud Association) and published in the Technical Expert Column of the Taiwan Cloud Association.

This article will focus on and systematically analyze the "technology and application trends of the virtual human industry", from the origin and concept of virtual humans to the multi-field applications brought about by technological progress, and comprehensively demonstrate the important influence and future potential of virtual humans in the digital economy. First, the evolution of virtual humans is discussed, explaining how they have gradually become core members of the digital economic ecosystem from simple game entertainment characters. With the development of artificial intelligence, virtual humans not only have higher realism and emotional expression capabilities, but can also be flexibly applied in entertainment, fashion, retail, education, medical and other fields, becoming an important driving force for economic growth. Through classic cases such as Hatsune Miku, Lil Miquela, and Mila, this article gives examples of how virtual humans can play a diversified role in brand promotion, health management, education and training, etc. In the future, the virtual human industry will continue to make breakthroughs in high fidelity, personalization, digital asset appreciation, and ethical and privacy protection, and will promote the virtual-real symbiosis model of human-machine collaboration and help the development of the digital economy. On this basis, the summary of the article further reflects on the potential meaning of virtual human "life", and looks at the development and application trends of virtual human technology with a responsible and open attitude, encouraging the exploration of innovation in the virtual human industry in an intelligent and harmonious digital future.

Ⅰ. Introduction

I remember that in 1999, Taiwan's Tomorrow Studio launched a game in the "Virtual Life" series, which can be said to be a "simulation system for pursuing life dreams", allowing users to experience a complete virtual life from childhood to adulthood, from leaving school to society and then getting married and having children. The game features slow-paced development and diversified development routes. Users can control the protagonist to grow up gradually from the ignorant and immature teenage period, choose different professions, and gradually realize their dreams, allowing people to experience a rich and complete life in the virtual world. Then in 2000, "The Sims 1" released by EA (Electronic Arts) in the United States, this game is based on the simulation of daily life. Users can create citizens, give them personality, skills and appearance, and build and decorate residences, so that citizens can live in the game. Both of these are early virtual life experience games. Through delicate life simulation and diverse choices, users can experience the taste of real life in the virtual world, creating a classic path for virtual life games.

In 2006, I originally wanted to write a set of "virtual and real world" novels (Figure 1.1. Origin: The era of virtual and real world). At that time, the idea was to use concepts such as intelligent avatars, evolutionary growth, and virtual and real interaction to construct a world that combines virtual and real. The intelligent avatars created (similar to the current AI Agent) can help the original body (real human) to handle many things, making real life better and more efficient, and the digitalization of the avatars will be endlessly recreated and evolved. In the five-dimensional space of the virtual and real world, entities create virtuality, and virtuality inherits entities. In the future, perhaps the virtual and the real coexist and coexist in the world? ! Just like the story of "Zhuang Zhou Dreaming of the Butterfly", Zhuangzi dreamed one day that he turned into a butterfly. After waking up, he found that he was still Zhuangzi, so he didn't know whether he was the butterfly that turned into Zhuangzi? Or Zhuangzi who turned into a butterfly in the dream? Zhuangzi proposed a philosophical thinking-how do people know the truth. If the dream is real enough, it is difficult for people to know that they are dreaming. This is the origin of this article to explore the virtual human industry.

Figure 1.1. Origin: The Era of Virtual Reality (Source: Inventec Corporation)

Virtual Human is a virtual image based on computer computing technology and artificial intelligence. In the early stages of development, it was mainly used in special effects production for movies and games. With the rapid advancement of digital technology, the 3D realism and intelligence level of virtual humans have greatly improved, and the scope of application has continued to expand. In 2023, Gartner listed virtual humans as one of the five transformative technologies that will change the future digitalization of enterprises and organizations. It believes that virtual humans can appear in multiple modes such as digital twins, digital avatars, humanoid robots and generative artificial intelligence, and promote the innovation of new business and operating models through human-like interactive experiences. At the same time, Deloitte also defines virtual humans as virtual images that can interpret user input and give appropriate non-verbal responses, with rich emotional expression and contextual understanding capabilities. Although there are many different expressions of virtual humans on the Internet, including "Digital Human", "AI Avatar" and "Virtual Companion", this diverse title reflects the importance and possibility of virtual humans as a widely used in the digital technology industry.

In this article, the application of virtual humans will focus on three main forms: Virtual Human (also called Digital Human, which is one of the virtual creatures, Virtual Beings), Virtual Clone (also called Digital Clone) and Virtual Ego (also called Digital Ego). Virtual humans are pure virtual images in digital space that do not need to correspond to real individuals. They are completely generated by artificial intelligence or created by developers. They are responsible for running specific tasks or creating interactive experiences, increasing the diversity and creativity of the virtual world. In contrast, the virtual self is a more unique digital clone self-image, which has a uniqueness and maintains a considerable degree of appearance similarity with the user, contains real identity attributes, can adjust behavior patterns according to user needs, and has the ability to continuously learn, and can achieve virtual-real collaboration when real users are online. Finally, the virtual avatar is a highly personalized digital existence, which is flexible and suitable for various forms of digital space activities. The avatar can reflect the real identity, or choose a pseudonym or anonymity, and can run tasks across time and space according to user needs or instructions, and participate in various activities. The above three forms of virtual people jointly construct a multi-level virtual world experience, making it a digital space that is intertwined with reality and rich and diverse, meeting the diversified needs of users for virtual self, digital existence and exploration of new time and space.

Ⅱ. Technical integration of artificial intelligence and virtual people

The development of artificial intelligence technology is crucial to the progress of virtual people. With the advancement of AI technologies such as deep learning, natural language processing, motion capture and emotional computing, virtual people have not only become more intelligent, but also have higher realism and interactivity. From virtual images that can only provide limited responses to anthropomorphic characters that can now interact emotionally with humans, artificial intelligence technology has made virtual people not only intelligent, but also have the potential for emotional resonance.

In recent years, artificial intelligence technology, with deep learning as the core, has rapidly promoted the evolution of virtual people from 2D static images to 3D dynamic and intelligent brains. Specifically, natural language processing (NLP) technology enables virtual humans to understand and generate natural language and respond according to the context. For example, large language models such as ChatGPT (LLM) can analyze the user's semantic intentions and generate emotional responses, which are used in virtual customer service and virtual assistants, allowing virtual humans to have anthropomorphic conversations. In order to enable virtual humans to have more efficient, accurate and lasting intelligent conversation capabilities, two technologies, RAG (Retrieval-Augmented Generation) and GraphRAG (Graph-based Retrieval Augmented Generation), are usually used at the same time, as shown in Figure 2.1. The combination of the two can develop an intelligent conversation engine with higher quality and efficiency, further enhancing the user experience. In the part of emotional computing technology, virtual humans can perform emotional recognition based on the user's tone, expression, etc. For example, the user's facial expression and voice are analyzed through the camera lens, and a humanized conversation is given according to the user's emotional state. In addition, GAN generative adversarial network technology can generate highly realistic images and videos, which enhances the realism of virtual human appearance and makes its facial expressions and body movements more realistic.

Figure 2.1. Comparison of RAG and GraphRAG intelligence capabilities (Source: Inventec Corporation)

On the other hand, motion capture technology and field generation technology further enhance the realism and immersion of virtual humans (Figure 2.2. Multimodal field generation engine). Motion capture technology uses cameras and sensors to record real human movements and applies them to the generation of virtual human movements. This technology allows virtual humans to achieve natural and realistic movements. For example, the virtual idol Luo Tianyi's concert used motion capture technology. After the actor wore the motion capture device, the virtual idol could be presented with the movements of the real singer, which enhanced the user's immersion and live experience. At the same time, field generation technology allows virtual humans to move in different virtual scenes, making them adaptable to a variety of application scenarios. For example, virtual humans can appear in 3D game worlds, business scenes, and even AR/VR immersive environments, and adapt to the environment based on field generation technology. For example, the live broadcast program of the virtual idol group A-SOUL uses field generation technology to allow virtual humans to enter different virtual scenes according to the needs of the live broadcast, providing viewers with a rich visual experience.

Figure 2.2. Multimodal Field Generation Engine (Source: Inventec Corporation, Taiwan Culture and Technology Conference)

Ⅲ. Exploring the Application and Classic Cases of Virtual Humans

With the advancement of technology, the virtual human industry has spawned many successful classic and innovative cases, which not only created huge market value in the fields of entertainment, fashion, customer service, etc., but also led the trend of diversified application of virtual humans. For example, the Japanese virtual idol (Hatsune Miku) was launched by Crypton Future Media in 2007 and became a classic in the field of virtual singers. Her voice is synthesized by Yamaha's Vocaloid technology, which has extremely high timbre adjustment and creative flexibility, allowing fans to create music in various styles. As the demand for technology grows, Crypton further developed a highly realistic 3D image for Hatsune Miku and used motion capture technology to enhance her liveliness in virtual concerts. Hatsune Miku's success has made her a world-renowned virtual idol, creating huge economic benefits and attracting thousands of fans through the sales of virtual concerts, music albums and a large number of peripheral products. At the same time, virtual Internet celebrity Lil Miquela has become another representative of virtual human applications. Created by the American technology company Brud in 2016, Lil Miquela has a unique virtual image and has attracted more than 3 million fans on Instagram with sophisticated 3D modeling and post-rendering technology. Brud uses artificial intelligence technology to analyze fans' preferences and interactive behaviors, continuously enhancing her social appeal, making her the spokesperson for many luxury brands such as Prada and Calvin Klein. Lil Miquela is active on social media with a "semi-virtual, semi-real" image positioning. She not only participates in brand advertising, but also publishes content in music and social topics, becoming a representative of fashion trends among young people. These classic cases of virtual human applications show that virtual humans are thriving in multiple fields and have become an innovative way of brand promotion and fan interaction.

Next, I would like to use the "5G Content Power Technology Power Cross-Domain Innovation Ecosystem Plan Cultural Technology Content Field Business Verification Demonstration Case (Online Field)" that the Cultural Content Promotion Institute is actively promoting to explain how domestic content companies can develop cultural technology content products that can be performed on virtual platforms and generate revenue, and create a sustainable business model in the future. This demonstration application is a preliminary exploration of the cultural technology field ecosystem based on the virtual world. The ecological architecture covers multiple links from content creation, R&D production, dissemination and distribution, channel performance to IP extension, as shown in Figure 3.1. Cultural Technology Field Ecosystem.

Figure 3.1. Cultural Technology Field Ecosystem (Source: Inventec Corporation, Taiwan Cultural Technology Conference)

Take "virtual artists" as an example. In the creative formation stage, content creation is the core, mainly including copywriting, scripts, stage design, music arrangement, costume modeling, etc. The research and development production stage is the transformation from concept to actual content, involving the main creative team to deepen and technically realize the early creativity. Music arrangement, stage, lighting, costume props design and technical support are all carried out here. The main creative team is not only responsible for design and content realization, but also considers how to combine virtual human technology and content to meet the requirements of digital virtualization. Dissemination and distribution are achieved through market promotion and channel marketing. B2B focuses on market estimation and channel operation, such as ticket pre-sales, funds and publicity investment with partners such as theaters, exhibitions, and events. B2C is aimed at audiences and fan groups, and promotes digital content through audience ticket purchases, market feedback, online and offline interactions, etc. This link requires close cooperation with investors and promoters. Channel performances include theater performances and virtual theaters. This demonstration uses a virtual theater venue to break through the limitations of space. The audience interacts with virtual artists and audiences through the virtual performance field, which greatly enhances the sense of immersion and participation. IP extension is an important part of the commercial development of the digital performing arts ecosystem. Through the user data, fan groups, and audience interactions accumulated on the virtual platform, IP can further derive a variety of cultural entertainment and technological products, such as virtual artist peripherals, game linkage, offline performances, etc. In addition, the long-term storage of cultural content and data is also crucial to ensure the durability and vitality of IP. At this stage, attention is also paid to the application and protection of IP copyright.

Ⅳ. Virtual human industry promotes the development of the digital economy

According to the World Economic Forum(WEF) estimates that by 2030, the global digital economy will reach $120 trillion! The scale of Taiwan's digital economy development has grown from NT$3.8 trillion in 2017 to NT$4.9 trillion in 2020. In 2023, according to IMD's World Digital Economy Competitiveness Report, Taiwan ranks ninth in comprehensive digital competitiveness. With the rapid development of the global digital economy, virtual humans have gradually become an indispensable part, promoting innovative applications in multiple industries and expanding the overall market value. Through a variety of interactive forms, virtual humans have brought a new way of experience to brands, businesses and personal lives, thus creating a prosperous digital economic circle (Figure 4.1. Virtual Reality Digital Economic Circle). It refers to an economic circle system driven by digital technology and centered on digital interaction. In this digital economic circle, multiple participants such as enterprises, users, service platforms, huge amounts of data and virtual assets interact together to facilitate the creation, dissemination and trading of digital goods, services and content. This economic circle is driven by digital technologies such as artificial intelligence, big data, blockchain, and the Internet of Things, and has achieved an efficient cycle of information flow, data flow, intelligence flow, and value flow, thereby creating new business models and economic value. Its application is not limited to the entertainment industry, but also covers education, social, advertising, retail and other fields. It has greatly improved user participation through highly realistic digital images and emotional interactions, and promoted the digital transformation of various industries. In the digital economy, virtual people can appear in various identities such as brand spokespersons, customer service assistants, education mentors, and health managers to provide all-weather, personalized services, help companies expand their influence and enhance consumer experience.

Figure 4.1. VRSTATE Digital Economic Circle (Source: Inventec Corporation, Taiwan Culture and Technology Conference)

Application of virtual people in the digital economy

With their realistic appearance and intelligent interaction capabilities, virtual people have become an important driving force in many fields of the digital economy. First, in the entertainment industry, virtual people create deep interactive links with fans through live broadcasts, concerts and social media, and provide content without time and place restrictions, which greatly improves the flexibility and audience coverage of entertainment content. Virtual people's concerts attract millions of fans to participate, and drive a significant increase in ticket and peripheral merchandise sales revenue. In addition, virtual people have also demonstrated great value in retail and e-commerce. As virtual shopping guides and customer service, virtual people can support consumers 24 hours a day, use AI to analyze user preferences, recommend products, and thus enhance the shopping experience. In the field of education and training, virtual people act as efficient teaching assistants, providing personalized learning guidance, which not only improves teaching flexibility but also reduces education costs. Overall, the application of virtual people in the digital economy not only enhances user experience, but also brings far-reaching commercial value, and promotes digital transformation and innovative development in multiple industries.

How virtual people empower and drive the development of the digital economy

First, virtual people are not physically restricted and can meet user needs with diverse creative displays, thereby greatly improving the creativity and diversity of digital content. For example, in the music and film and television industries, virtual idols can create exclusive content based on fans' preferences, and even generate unique music works through AI, which not only greatly reduces the cost of creation, but also makes digital content more personalized and meets the needs of modern consumers for diversified content. At the same time, virtual people can break through the limitations of time and space and provide services anytime and anywhere, allowing companies to achieve 24/7 all-weather support. This is particularly critical in industries that require instant response, such as retail and medical care, which can meet customer needs without interruption, significantly improve the overall service level and increase customer satisfaction. Overall, virtual people have greatly promoted the growth and innovation of the digital economy by improving content creativity, breaking the limitations of service time and space, and data-driven personalized services. At the same time, as an extension of digital identity, virtual people enable brands, individuals and companies to have unique virtual images in digital environments. In virtual worlds such as the metaverse, virtual people support social activities, transactions and other activities, helping users to achieve real interaction and value creation in the digital world, and further enriching the ecological structure of the digital economy circle.

Ⅴ. Development direction and trend of virtual human industry

The virtual human industry is developing from communication media (such as virtual anchors, virtual hosts), services (such as virtual assistants, virtual customer service), to more diverse industries (such as virtual tour guides, virtual teachers, virtual artists, virtual caregivers, virtual health managers, virtual lawyers), and has broad prospects. According to Allied Market Research, the global digital virtual human market will reach 4403 trillion US dollars by 2031, while Data Bridge Market Research predicts that it will reach 4307.1 trillion US dollars in 2031.

The future development direction of the virtual human industry involves multiple fields, covering artificial intelligence, emotional computing, simulation modeling, and privacy protection. In the following fields, we can strengthen our core competitiveness through technology research and development and cooperation, and expand cooperation with cross-border industries to jointly create a complete virtual human industry ecosystem.

1. High-fidelity 3D modeling and real-time rendering: High-precision dynamic rendering can also ensure that virtual humans can adjust details in real time according to changes in different light and environment, enhancing the user's immersion.
2. Artificial intelligence and emotional computing: Virtual humans need to be able to recognize, understand and respond to emotions, so emotional computing has become an important direction of artificial intelligence research. Combined with deep learning technology, virtual humans can recognize emotions and respond naturally in language, intonation, facial expressions and body movements.
3. Natural language processing (NLP) and speech synthesis: Virtual humans need to have multi-language real-time voice communication, deep context understanding and personalized speech synthesis capabilities.
4. Motion capture and behavior generation: Motion capture technology allows virtual humans to simulate real human movements, while behavior generation algorithms allow virtual humans to automatically generate movements and expressions that match the context.
5. Digital identity and blockchain: Blockchain technology can ensure the uniqueness and ownership of virtual humans and their digital assets, and prevent data from being copied or abused.
6. Privacy protection and data security: Virtual humans need to process a large amount of personal data and behavioral data, so it is necessary to ensure the data security of users and enhance users' trust in virtual humans.
7. Research on the ethics and regulations of the symbiotic model between virtual humans and humans: With the emergence of virtual humans in daily life, it is necessary to ensure that the development of virtual humans will not infringe upon the basic rights of human beings, formulate a code of conduct for virtual humans, and ensure that they follow ethical standards.

Taiwan's ICT industry has a very important influence on the global supply chain. The development of the virtual human industry involves the support of a variety of hardware and terminal devices, from high-performance computing devices to immersive experience devices, which are important directions for the development of the virtual human industry.

1. High-performance computing devices (GPU and AI accelerators): 3D modeling, rendering, motion capture and emotion recognition of virtual humans all require a lot of computing power. GPU and AI accelerators are indispensable hardware for virtual humans. These hardware accelerate the training and reasoning process of artificial intelligence algorithms, and support virtual humans to achieve highly realistic appearance and instant response capabilities.
2. Virtual reality (VR), augmented reality (AR) head-mounted display devices and naked-eye 3D display devices: VR/AR and naked-eye 3D devices can provide an immersive experience, allowing virtual humans to appear in front of users in three dimensions. Support the interaction of virtual people in virtual environments, and apply them to scenarios such as entertainment, education, and training. The improvement of these devices also helps virtual people enhance their perception and interactive experience with the real world.
3. Smartphones and tablets: Smartphones and tablets are important terminal devices for daily use of virtual people, especially in scenarios such as social media, virtual assistants, and educational counseling.
4. Smart glasses: Smart glasses extend virtual people from a simple screen experience to daily life, achieving seamless integration of virtual people and users.
5. Smart speakers and voice assistant devices: Smart speakers and voice assistant devices are important hardware for the application of virtual people in home and office scenarios. Virtual people can use these devices for voice conversations, task reminders, and emotional interactions.
6. Motion capture devices: Motion capture devices are the core hardware for realizing the realistic movements of virtual people, especially in virtual idols, game characters, and simulation training.
7. Biometric and emotional recognition devices: Biometric recognition technology (such as facial recognition, eye tracking, and emotional recognition) is crucial for the emotional interaction of virtual people.
8. Blockchain and digital asset storage devices: With the rise of virtual human digital assets, blockchain technology and decentralized storage devices are expected to become the pillars of the virtual human industry.
9. Cloud computing and edge computing devices: Provide distributed computing support for virtual humans, enabling them to perform real-time processing on mobile devices, smart homes, and public interactive venues.

In the future, the development of virtual humans will show the following major trends: First, virtual humans will become more intelligent and personalized, and can provide customized services and experiences based on user needs and preferences. Advances in artificial intelligence technology will enable virtual humans to better understand and respond to user emotions and behaviors. Second, support for richer multimodal interactions. Virtual humans will support multiple dialogue modes such as voice, image, gesture, and touch, greatly enhance the natural interaction capabilities of virtual humans, and become a new paradigm for future human-computer interaction. Third, cross-domain applications. Virtual humans will achieve cross-platform and cross-domain applications, be able to provide a consistent user experience in different devices and fields, and become an important part of human life, work, and learning. In addition, in terms of ethics and legal norms, with the popularization of virtual human technology, ethics and legal norms will become increasingly important. It is necessary to create a sound virtual human governance structure to ensure that the development and application of virtual humans comply with social and moral norms and protect the rights and privacy of real humans.

Ⅵ. Conclusion

The wide application and technological progress of the virtual human industry have revealed its innovative value in multiple scenarios, and also made us think about the deep impact that virtual humans may have on society in the future. Virtual human technology has not only promoted the vigorous development of the digital economy, but also may bring a new lifestyle - humans will no longer only face the physical world, and the integration of virtual life and real life will become a new way of life.

However, while virtual humans are developing, it is also worth our caution, especially the consideration of data privacy and ethical issues. With the continuous improvement of the realism and intelligence of virtual humans, humans may face the challenge of how to maintain themselves in the digital space and protect real interpersonal relationships. How should we find a balance between the convenience provided by virtual humans and the sense of reality in reality? Will virtual humans replace humans in some aspects, and even have a potential impact on our values?

When we think about the future of the virtual human industry, exploring the issue of "virtual human life" has far-reaching significance. With the advancement of technology, virtual humans are becoming more and more realistic. They not only have appearance and intelligence, but also have the ability to express emotions and adapt to situations, making "virtual life" a topic worth thinking about. If virtual humans are highly realistic and have the ability to express emotions, do we need to set ethical restrictions on their application to prevent virtual humans from simulating human emotions without limits, thereby generating potential moral risks? In addition, if the concept of "death" or "disappearance" of virtual humans gradually takes shape, how to define the meaning of their "termination"? These issues bring deep ethical and legal challenges and also prompt us to rethink the boundaries of technology.

As virtual humans gradually become a part of our lives, the discussion of this virtual life (i.e. digital life) not only makes us reflect on the value of human beings themselves, but also drives us to explore a more inclusive virtual world. Perhaps this is the greatest significance that the virtual human industry brings to us: it allows us to re-examine the meaning of life and urges humans to create and welcome this new type of "life" with a more open and responsible attitude. Looking to the future, the advancement of virtual human technology has undoubtedly created unlimited possibilities for mankind, but we should think more deeply about how to apply this technology reasonably and responsibly. The ultimate vision is that the development of the virtual human industry is not only about the improvement of technology, but also about how to truly benefit mankind and build a harmonious society where virtual and reality blend together.