Three years of empty fire cannot create a future for virtual people

In recent years, virtual humans have appeared in the public eye many times, but they have not left a deep impression. Is it because of insufficient technology? Or is it because of the single content? Or is it because of the limited application scenarios? What is the future of virtual humans? The author of this article analyzes and answers the above questions one by one. Let's take a look.

In recent years, virtual people have gradually broken through the barriers of the second dimension and started to appear in mainstream media with real images, attracting the attention and pursuit of young people.

At the beginning of this year, virtual idol Luo Tianyi once again appeared on the New Year's Eve party of Bilibili, becoming the hot topic of the evening party. Later, in the promotional video of CCTV's Lantern Festival Gala, the virtual person "CCTV Net Xiao C" made his debut, bringing the audience scenes and content rich in Chinese classical culture. For a time, the image of virtual people once again became a hot topic among netizens.

Looking back on the past year, a large number of virtual human images with their own characteristics, such as "Wen Yaoyao" from the China Cultural Relics Exchange Center, "Xiaozhu" from Xinhuanet, "Gu Xiaoyu" from Zhejiang Satellite TV, and "Jia Yao" from the Dunhuang Research Institute, have appeared in front of C-end users, and the application scenarios are becoming more and more abundant.

The picture shows Xiao C from CCTV.com

However, after several years of drastic technological innovation, the current image of virtual humans does not seem to be significantly different from the early days. Applications are still mainly concentrated in the technical services of the B-side and the content entertainment field of the C-side, and there is insufficient profit margin. As domestic and foreign large companies continue to lay off and adjust their metaverse business lines, the research and development and investment of various companies in virtual humans are also shrinking. All these signs inevitably make people wonder - has the future virtual human track reached a bottleneck and to what extent can it develop? ?

From technology to content, what difficulties has the development of the virtual human industry encountered? What is the future development direction and the way out?

1. Technology makes virtual humans popular

What is a virtual person?

At present, the definition of virtual humans that is generally accepted by the industry is the explanation in the "2020 White Paper on the Development of Virtual Digital Humans" released by the China Artificial Intelligence Industry Development Alliance: Virtual humans refer to virtual characters with digital appearance, who rely on display devices to exist and have human appearance, human behavior and human thoughts.

In essence, virtual humans are virtual or simulated characters created through computer graphics rendering, motion capture, deep learning, speech synthesis and other technologies. Since 2020, with the accelerated entry of capital into the digital human market, it has promoted the rapid development of technology and the accelerated exploration of application scenarios, bringing a lot of application space.

From the perspective of application, the virtual digital humans that have been born can be generally divided into two categories: one is a service-oriented virtual human that focuses on functional applications, and the other is an idol-type virtual human that focuses on personal expression.

Currently, service-oriented virtual people mainly replace some repetitive manual labor, such as intelligent customer service, voice robots, virtual assistants, etc. In addition to standardization and regularization, their value lies in providing real-time services around the clock and improving work efficiency. For example, Vanke's first virtual employee "Cui Xiaopan" is responsible for urging the company's prepaid receivables overdue bills, and the museum's virtual guide, etc.

Pictured: Cui Xiaopan

Idol-type virtual people can meet the value needs of brands by creating unique personalities that meet the aesthetic tastes of contemporary young people. The identity symbols of this type of virtual people are mainly virtual anchors, brand spokespersons, fashion experience officers, etc. They are the most widely exposed to users and the fastest commercialized type, mainly focusing on marketing and content fields.

The development of virtual humans did not happen overnight. In the past 20 years, virtual humans have become a technology-driven emerging industry as computing space and computing platforms have been continuously upgraded.

From the perspective of technological updates and iterations, the development of virtual reality has gone through three important stages:

The first is the "paper man" stage, where game or animation companies usually release virtual images and use electronic sound synthesis software to make the virtual images sing and dance. At this stage, users can only watch but cannot interact with them.

The second stage is human-driven. In this stage, technicians not only use software to make fixed arrangements, but also use motion capture equipment to drive virtual humans to perform and interact in real time.

The last stage is the autonomous interaction between virtual humans and humans. Currently, through artificial intelligence technology, some virtual humans can understand external content and provide real-time feedback.

Thanks to the rapid development of a number of basic technologies, the application scenarios of virtual humans have been continuously expanded, the market scale has been rapidly expanded, and a large number of companies have flocked in. During the three years of rapid development of the industry, many new and diversified ways of playing have emerged in the field of virtual humans.

2. Review of the past three years: Diversity of virtual humans

At present, the technology of virtual digital humans is gradually maturing, and the industrial chain has taken shape. Many companies, especially major Internet companies, are actively making plans and creating a large number of application scenarios for virtual humans.

For example, Baidu’s digital human live broadcast platform “Xi Ling” integrates digital human production, content creation, and business configuration services, and provides creation and operation services for virtual hosts, virtual employees, virtual idols, and brand spokespersons. It is mainly used in live broadcasting and enterprise marketing. ByteDance’s digital human layout is mainly in the virtual idol business A-SOUL, and forms a linkage with PICO. Tencent’s virtual human layout is centered around the “Xiaowei Digital Human Platform”, which is mainly used in B-side scenarios, including the broadcasting digital human platform and the interactive digital human platform.

Looking back at the development of virtual humans, the gaming field is considered to be the earliest application scenario. Virtual humans are mainly used in the gaming field to create "digital avatars" for players. For example, the already mature face-pinching technology in games essentially creates a virtual human image for players and opens up the customization rights of the virtual human's appearance, clothing and other elements to players.

At the same time, based on the ability of virtual reality technology to simulate and connect the real world, major domestic and foreign manufacturers have begun to try to create virtual human images for players. In the future, players will be able to have a virtual image in the game that matches their actual image and will be able to automatically switch perspectives, allowing players to obtain a perceptual experience similar to the real world in the virtual world.

In addition to the gaming sector, virtual idols have also brought huge commercial space to the field of virtual people. According to data from relevant institutions, more than 80% of Internet users in my country have the habit or willingness to chase stars, of which 63.6% of users will support and pay attention to the relevant dynamics of virtual idols, and 37.6% of users are willing to pay to support virtual idols. Japan's "Hatsune Miku", China's "Luo Tianyi"... Some virtual idols that have successfully gone viral have begun to be active in mainstream media and even appeared on the stage of the Spring Festival Gala. Their communication and influence are no less than that of real stars. The virtual idol culture has begun to explode.

Another active scene for virtual idols is in short videos. Virtual images in short videos often appear in an exquisite virtual scene, building a cool content universe and exploring new ways of presenting short video content through technology. The most representative one is the beauty and makeup expert "Liu Yexi" who can catch monsters. One of her videos contains a variety of popular content categories such as suspense, plot, beauty and post-production special effects. Her first video gained millions of fans. For a time, various secondary creations such as imitation makeup and cosplay around Liu Yexi broke out on the platform, and she once became the most talked-about virtual image at the time.

The popularity and topicality of virtual people have brought space for the commercial development of virtual people. Many brands have discovered the marketing advantages and value of virtual people and used virtual people as an innovative means to reach young consumers. The development of virtual people is largely in line with the laws of the fan economy, which has gradually attracted major brands to virtual images with positive energy labels and a large audience of Generation Z, and reached cooperation with them. In addition, some brands choose to develop and create their own virtual human IPs. Colonel KI of KFC, the virtual image of Hua Xizi with the same name, Watsons' virtual spokesperson Qu Chenxi, etc. are all developed by brands, and have stronger marketing and advertising attributes than other virtual people.

The picture shows the virtual images of Huaxizi and KFC

From content value to commercial value, the development of virtual humans has shown us the driving force of rapid development. However, the industry, which should have been on the rise, has gradually become quiet in the three years of exploration and has entered a stage of slow progress. The main reason is that the bottleneck of hardware technology is difficult to break through and insufficient attention is paid to the construction of content ecology.

3. The “false fire” is gradually extinguished, where lies the problem with virtual humans?

The underlying reason for the slowdown in the development of virtual humans is the immaturity of hardware technology.

Due to high costs and insufficient functions, hyper-realistic virtual humans have not yet been widely used. Usually, the production of a virtual human requires the integration of several companies to complete the entire chain. The character setting, role creation, role action, and engine drive of the virtual human are distributed in different production companies. The various links are relatively fragmented, and the cost of producing highly intelligent and interactive virtual humans is relatively high. According to industry insiders, the current cost of producing a virtual human with AI-driven capabilities is at least one million yuan, of which more than half is technical investment. Under such circumstances, the income of many companies cannot cover the costs, which ultimately leads to insufficient technical investment.

From the user's perspective, the current operation of virtual images mainly relies on technologies such as motion capture and portrait simulation, and focuses more on external representations such as movements and expressions to make virtual people look like real people. However, these technologies cannot give virtual people subjective thinking and emotions, such as the ability to perceive emotions and empathy, which makes it difficult for users to break away from the inherent impression that "virtual images are fake."

At the same time, in order to make the movements of virtual people more realistic and natural, virtual images are often deeply bound to real people, and even gestures and speech must be completed by portrait capture and dubbing. This leads to the situation that when the virtual people interact with users on stage, if the "person in the middle" behind them moves too much or the technology is unstable, it is easy to cause accidents such as penetration and loss of voice. For example, when Luo Tianyi was singing in Li Jiaqi's live broadcast room, she lost her voice on stage due to a technical failure, which greatly reduced the authenticity of user interaction and the viewing experience of the live broadcast.

Not only that, the "person behind the virtual person" cannot be easily changed. The words and deeds of the virtual image must be consistent, otherwise there will be a risk of "personality collapse". The most well-known incident is the Japanese virtual idol "Kizun Ai". The company behind it wanted to operate the virtual image as a "dopant". It successively let four people play the image of "Kizun Ai", which caused strong dissatisfaction among fans. Many of them quit and turned against her, resulting in a large-scale loss of fans.

In addition to hardware technology, another key difficulty in the current development of virtual humans is that there is no deep integration between virtual humans themselves and their application scenarios.

Taking the combination of content scenarios as an example, there are three main operation paths for virtual people at present: the first is brand and advertising marketing, marketing and endorsement on channels such as Xiaohongshu and Weibo; the second is entertainment live broadcast, making profits through fan rewards; the third is combining virtual people with games, short videos, etc. However, the application of virtual people in these scenarios is more superficial, or just exists as a gimmick.

From the perspective of live broadcasting, Bilibili has launched a series of virtual image anchors such as Zuya Naxi, Duoduo Poi, and Mocha Official, which attracted the attention and support of many users in the early stage of live broadcasting. However, the live broadcasting content and form of these virtual anchors are very simple, that is, they appear in virtual images and interact with users through dubbing. There is no clear content difference. After users become aesthetically fatigued, their popularity quickly declined.

In addition, virtual people such as Hatsune Miku and Ling Ling have appeared in live broadcasts with goods many times. The comments in the live broadcast room are more about whether they are real people, whether they can sing and dance, etc. The main focus of users is more on the virtual people themselves, and issues unrelated to the goods. Virtual people themselves often do not introduce too much information about the product, but are more about providing entertainment. Therefore, there is no deep match between the virtual people and the live broadcast room, and they are more like a "vase".

Image source: Screenshots of some virtual people’s live broadcast rooms

In the final analysis, the content scenarios of virtual people have not been effectively connected, the content environment for in-depth creation of virtual people has not yet been formed, and they do not have the business capabilities required by customers, making it difficult to match monetization with investment. At the same time, the virtual people that have been born are limited by technology and lack personality, making it difficult to communicate emotionally with users.

4. The future is still “future”, can the half-open door to commercialization be opened?

It is precisely because of the various constraints that virtual humans are facing today that the overall commercial development space of the industry has not been fully developed. However, although the current profits of the virtual human market are lower than expected, the growth trend of domestic virtual human-related companies shows that the "cake" of this track is still tempting.

According to Qichacha data, the number of related enterprise registrations has shown an overall upward trend in the past five years. In 2021, more than 1,000 related enterprises were newly registered, a year-on-year increase of 63.69%. In the first quarter of 2022, nearly 300 related enterprises were newly registered, a year-on-year increase of 67.47%, of which 11 enterprises have obtained financing.

The relatively mature monetization methods of virtual images are mainly to provide marketing empowerment for brands and monetize around the fan economy. However, with the continuous increase in the size of the virtual human market and the purchasing power of Generation Z fans, virtual images have gradually jumped out of the scope of niche culture and have huge room for development in terms of commercial value.

As mentioned above, the monetization of virtual people is mainly based on the fan economy. The prevalence of the fan economy has attracted major brand owners to virtual images with positive energy labels and a huge audience of Generation Z. Major brands have injected a sense of the future and technology into their brands through virtual images to attract new fans from different circles.

For example, Zhixuan teamed up with Qiu Jirong, the direct descendant of the Qiu School of Peking Opera, and the popular virtual idol Ling_Ling to launch the film "Winter Olympics Forbidden City Peak", which combines traditional culture, sports events and brand concepts. After the film was launched, the brand not only reached users who pay attention to the Winter Olympics, but also users in the fields of ancient style and the second dimension expressed their love for the film. Through virtual images, cross-circle communication of brands has become an interesting and novel thing.

Image source: Ling_Ling Weibo screenshot

Not only that, as ChatGPT has taken up the "banner" of AI development, the combination of virtual humans and ChatGPT has also brought new monetization space to virtual humans. Just in February of this year, companies that designed virtual human-related businesses, such as Shiyou Technology and Fengyuzhu, announced that they were developing related technologies to connect virtual humans to ChatGPT. Even the news that Baidu is about to launch the Chinese version of the "GPT Big Model" Wenxin Yiyan was officially announced and released by Baidu's virtual humans such as Xijiajia and Du Xiaoxiao. The combination of virtual humans and ChatGPT may make generative AI no longer rely solely on text, but output in the form of vivid virtual humans. In the future, the diversity of presentation methods will bring it more commercial gameplay.

When virtual people are able to actively create content and provide real-time feedback, they will present visuals in a more personalized and interactive way, thereby stimulating the audience's senses and desire to express themselves, and upgrading the audience's emotional experience in the interaction with the virtual image. Whether it is entertainment live broadcast or bringing goods, it can touch users emotionally, thus forming conversions and profits in more interesting interactions.

It is worth mentioning that the existing virtual images in China often have marketing attributes, and there are very few IP images that rely purely on entertainment live broadcast content for monetization, which means that there is still room for development in the field of virtual live broadcast. For relevant institutions, grasping the traffic advantage of fan economy and seizing the entertainment live broadcast track on the basis of ensuring the integrity of virtual images may be able to take the lead in occupying the market share of virtual live broadcast.

From the macro perspective of the metaverse industry, in the future, in the further development of digital twins, not only will the entities and processes in the industrial field be able to provide digital twins in the virtual space, but more individuals will also be able to establish their own digital twins through virtual human technology, and the application scenarios of virtual digital humans will be further expanded. With the evolution of the "virtual-real symbiosis" stage, real users and their own virtual humans will form new social relationships and emotional connections, and complete a series of economic behaviors that can be carried out in the real society, which brings infinite imagination to the future value of virtual humans.

5. Final Thoughts

In the final analysis, the development of virtual humans and their industries has made significant progress, but is still in its early stages of development due to constraints such as technology, cost, and application scenarios. The calmness at this moment does not represent the decline of the industry, but rather the calmness and stability after the rapid explosion of the concept.

Driven by technology, virtual humans have gradually expanded from the initial virtual idols to virtual assistants and virtual anchors, and their forms are becoming increasingly diversified. With the continuous improvement of content presentation methods, virtual humans will accelerate commercialization under the wave of modern consumerist fan culture, feed back the innovation and research and development of virtual humans, and finally form a complete technology and content ecosystem.

For relevant institutions, virtual people with cultural connotations and significant differentiation are more likely to enter the public eye and become the "newcomers" in the content and marketing markets, while those virtual images with similar images and single content will eventually be eliminated by the market.

Author: Dake; Editor: Ji Nan

Source: TopKlout