In the past two years, various avatars have sprung up, and they have begun to take on tasks that originally required humans to complete, some providing services like humans, and some taking to the stage to perform. The wave of heat brought by the concept of metaversity has made a large number of enterprises and capital invest in the construction of avatars, but the most basic problem of "where to come from and where to go" has yet to be solved.
Baidu, which has invested heavily in artificial intelligence technology, has put forward its own new insights into avatars and metacosms. Baidu believes that this kind of image can be collectively referred to as "digital people", there are three major difficulties of industrial chain fragmentation, service scenes and performing arts scenes are not effectively opened, and high costs to meet high-frequency demand, and practical methods are needed to help enterprises complete the cost reduction and efficiency of service and performing arts production.
Based on this, Li Shiyan, head of Baidu Intelligent Cloud AI Human-Computer Interaction Lab, introduced the Baidu Intelligent Cloud Xiling Platform. This is a platform-level product integrating digital human production, content creation, and business configuration services, integrating Baidu's artificial intelligence technology achievements and digital asset services, so that "digital people" are expected to shift from high-cost customization to standardized and rapid delivery.
In baidu's context, "digital people" can be divided into three stages: the "paper person" of flat image + speech synthesis is 1.0, the virtual anchor can be regarded as 2.0, and the AI involved in modeling and content production is 3.0. Baidu now has the ability to produce "service-oriented digital people" and "performing arts digital people" in the 3.0 stage, and the Baidu Intelligent Yunxiling Platform can quickly generate 2D avatars through photos, and can also greatly compress the production time of 3D virtual images.
From the current product point of view, the concept of "service-oriented digital person" is more like a human-computer interaction interface that is close to the real world interaction mode, eliminating the strangeness of using electronic products through an anthropomorphic image, so that people who are relatively unfamiliar with digital literacy can easily use various services. Similar ideas first appeared at the end of the last century, but were eventually abandoned because it was difficult to integrate complex scenarios and understand requirements.
Nowadays, the re-proposed anthropomorphic human-computer interaction reform may be closely related to the rapid development of artificial intelligence and hardware products: products such as Baidu's large-scale online cloud customer service can provide users with a more natural way of interaction, and at the same time can more accurately understand feedback, while multimodal interactions such as images and voices have a basis for providing adequate services.
As for whether the "service digital person" will truly meet the needs of the metaverse world, I am afraid that it is not the best time to make a judgment. The meta-universe hardware and software technology foundation that can support rich details, stability and reliability, and acceptable to the public has not yet arrived, and such products must first provide interactive experience changes on existing terminals, such as in-web dialogue, mobile App operation or ATM.
The "showbiz digital person" provides another kind of imagination space, which looks more like a real person than the graphic and text, and communicates with users in a more secure way on behalf of brands, enterprises, projects. When it comes to entertainment consumption, social companionship and other needs, the virtual image has a "perfection" that real people cannot reach, and it can not change with time.
When computing resources explode and the cost of use becomes quite low, it may be possible to customize a unique avatar for each person, "all actions are only to meet your preferences." Li Shiyan demonstrated his vision: "Our goal in the past two years is to let everyone realize the freedom of digital people through the opening of Baidu's intelligent Cloud Xiling platform. ”
Of course, most of the current products are in a fairly early stage, and many avatars only meet the requirements of moving, and have not yet been able to impress people's hearts like real idols. However, the ancients had a cloud "do not accumulate steps, can not go thousands of miles", in the road to provide more diversified entertainment consumer products, Baidu intelligent cloud Xiling-driven "performing arts digital people", may be the first key step.