It is understood that Alibaba (BABA) Tongyi announced the open source release of LHM, which can drive the generation model of hyper-realistic 3D digital human, and can generate hyper-realistic 3D digital human in seconds with a single image.
It is reported that you only need to input a picture to have a low-latency real-time conversation with the digital human avatar generated by this picture. In the future, LHM has three major application directions: action reproduction, game character generation and virtual reality exploration.
AI helps digital human industry upgrade
Today, the application scenarios of digital human are very rich, especially in the fields of communication, education, e-commerce, etc. For example, in the live broadcast room, the digital anchor can interact in real time like a real person, 7×24 hours online, and the digital human live broadcast room can still “stick to the post” in the case of “unmanned driving”.
The development of digital human is much faster than expected, and AI empowerment is just a microcosm of its transformation. With the maturity of technology and the improvement of ecology, AI digital human will evolve from “functional assistant” to “emotional partner”, reshaping the paradigm of human-computer interaction.
In addition, the continuous upgrading of AI technology continues to expand the “working ability” of digital people. Compared with digital people driven by humans (virtual anchors, virtual idols, etc.), digital people driven by AI (AI assistants, AI digital employees) are now more popular after accessing multimodal large models.
In the current consumer field, AI technology has brought innovation to the animation production and entertainment industries. Creators can use AI virtual avatars to quickly generate dynamic characters, which can be applied to short videos, games or virtual anchor scenes, greatly reducing the production threshold and time cost.
In this regard, industry insiders pointed out that according to technical analysis, the ability of AI models in image generation and speech synthesis has greatly improved recently, and the emergence of AI virtual people and virtual avatars is the crystallization of the fusion of these technologies. With the help of advanced algorithms, AI virtual avatars can generate realistic facial expressions and natural and fluent speech, and even achieve accurate synchronization of lip shape and sound.
It is undeniable that the rise of AI digital people is not only a leap forward in AI technology, but also an important evolution of human-computer interaction in the digital age. At the application level, AI digital people have shown diverse potential. According to the “China Digital Human Development Report (2024)”, it is estimated that in 2025, the market size of the industry driven by Chinese digital humans will reach 640.27 billion yuan, and AI digital humans will move from concept to large-scale application.
WiMi promotes the development of AI digital humans
Obviously, AI digital humans are not only the product of technological innovation, but also the core tool for the intelligent transformation of enterprises. For enterprises, the layout of digital humans is not only a choice to seize the market opportunity, but also a must for future competition. Public information shows that WiMi (WIMI), as a full-stack technology provider for virtual digital humans, has been focusing on the research and development of full-stack technology and full-scenario applications for virtual humans for several years, helping digital humans to land in multiple industry scenarios such as radio and television media, government affairs, cultural tourism, education, and commercial consumption, and becoming an important boost to the intelligent transformation of the digital human industry.
So far, WiMi has achieved low-cost, short-cycle, and batch production of AI virtual digital humans, as well as low-latency, high-precision, and intelligent interactive experiences. Behind the creation of these numerous digital human images with both form and spirit, it is the strong support from the enterprise’s AI multimodal technology, and is based on the innovation of the AI virtual digital human generation platform. At the same time, a total of 4,600+ digital human IPs have been created, and the core business system based on “AI digital human creation + high-definition virtual broadcasting” has been continuously expanded to fully promote the commercialization of AI virtual digital humans.
Conclusion
As one of the first applications to be commercialized in the AI big model, AI digital humans are like a portal for human-computer interaction in the eyes of most people, and now they can communicate with humans without barriers. It can be foreseen that with the further iteration of generative models and the growth of market demand, AI digital humans may develop into physical digital humans with high intelligence and simulation interaction capabilities in the future, which will become an important part of the future technology landscape and bring more possibilities for the integration of creativity and business.
