In July, Robin Lee predicted at the Baidu World Conference that "AIGC (Artificial Intelligence Self produced Content) will subvert the existing content production model in the next decade. It can generate AI original content at a cost of one tenth and a production speed of one hundred times and one thousand times."

In August, in the Emerging Digital Artists Competition held in Colorado, the contestants without painting foundation submitted the AIGC painting "Space Opera House", but they won the first prize in the category of "digital art/digital retouching photos" in the competition, causing a storm in the industry.

Accelerated entry of capital and institutions: StabilityAI announced that it had received seed round financing of US $101 million and was officially promoted to Unicorn Bank; Almost at the same time, Jasper, another artificial intelligence unicorn company, also announced that it had obtained a new financing of 125 million dollars. For a while, there was a global AIGC investment boom.

| What is AIGC and what is its use?

AIGC refers to the automatic generation of content by artificial intelligence, which is considered as a new content creation method after professional production content (PGC) and user production content (UGC). At present, it is mainly used in text, images, videos, audio, games and virtual people, specifically:

1) Writing.

At present, AIGC generated text is mainly used for news writing, given format writing and style rewriting. For example, users can enter a description or requirement for the target article, and the system will automatically grab the data and create according to the instructions we describe.

2) Image creation.

The technical platform lowers the threshold of artistic painting creation. Users only need to input text descriptions, and the computer will automatically generate a work.

3) Video Creation.

For example, Google has introduced the AI video generation model Phenaki, which can generate a variable length video based on text content. In the published DEMO, Phenaki forms a logically coherent video based on hundreds of words in two minutes.

4) Audio clip.

The audio generated by AIGC has long been used in our daily life, such as the voice in the commonly used mobile phone navigation. The deeper application will be in the field of virtual human. AIGC can not only generate the voice of virtual human, but also create the content.

5) Game development.

The application of AIGC in games can be divided into two aspects. On the one hand, it is used to build scenes and stories. On the other hand, players can create their own virtual people through AIGC's platform tools, which can be used in games such as gold making.

6) Code generation.

According to the data, AIGC will develop at an amazing speed in 2022, and its iteration speed will explode exponentially. The continuous improvement of deep learning model, the promotion of open source model, and the possibility of commercialization of large model exploration will become the "acceleration" of AIGC development.

In terms of market space, Gartner predicts that by 2025, generative AI will account for 10% of all generated data. According to the analysis of Generative AI: ACreative NewWorld, AIGC has the potential to generate trillions of dollars in economic value. Guotai Jun'an said that in the next five years, 10% - 30% of the image content will be generated by AI, and the corresponding market size will be more than 60 billion.

| Where will the AIGC application scenario be?

The core value of any new technology is the innovative application under the scenario, and AIGC is no exception. Scenarios are an environment for technology application and an external condition for testing technology capabilities. The better the technology, the more universal the scenario application. Therefore, the scenario is an important factor to test the development status and future of AIGC.

At present, AIGC application scenarios in a broad sense can be roughly divided into two tracks: ToB and ToC. The application on the B side is mainly to enable various industries through "AI+", while the application on the C side is mainly to produce content through interaction with people, mainly including AI driving, AI assistant, AI auxiliary text generation, etc.

In July this year, technology maniac Mask announced that he had successfully uploaded his brain to the cloud, which not only realized human-computer dialogue, but also permanently saved the brain data. According to his idea, with the continuous development of AIGC technology, the mapping from the real world to the digital world can be established.

| Technology giants begin to layout

Domestic giants Baidu, Tencent Youtu, Alibaba, Kwai, ByteDance, NetEase, Shangtang, and Meitu have all invested in the AIGC field.

Robin Lee judged that AIGC will usher in three stages of development: "assistant stage", where AIGC assists humans in content production; In the "cooperation stage", AIGC appears as a virtual human with virtual reality, forming a human-computer symbiosis situation; In the "original stage", AIGC will independently complete the content creation.

The manuscript writing robot "Dream Writer" created by Tencent; Alibaba's AI online design platform Lubanner helps marketers produce Banners; ByteDance's clip image and Kwai cloud clip can provide AI generated video; The one-stop AI music creation platform "Netease Skytone" launched by Netease, etc.

In the field of AIGC, foreign countries fight with immortals. The technology giants Google, Meta, Microsoft, etc. also have some new unicorns from AIGC, such as StabilityAI, Jasper, OpenAI, etc.

Technology companies soon extended the popularity of AI painting to AI video generation. From Meta's announcement of Make-A-Video, a text to video system, to Google's announcement of ImagenVideo and Phenaki, which can generate high-definition video from simple text prompts, AIGC has developed rapidly overseas.

| Is it opportunity or foam?

The essence of digital content+unique originality determine that the capacity of AIGC's track is large enough. The former implies that it can be built into a standard product like Internet content platform or product, and the latter means that it can obtain market recognition comparable to that of human creators.

Although AIGC is currently in a state of explosion, in the final analysis, the enterprise's development of new technology depends on whether the technology has commercial value, while investing in an enterprise depends on whether the core technology can make it continuously profitable.

High initial investment requires the rapid expansion of AIGC enterprise user scale. Because only when the number of users in the industry reaches a certain size can the cost be amortized and the loss be turned into profit. Past experience shows that to achieve rapid growth of user scale in a short period of time, it is necessary to provide standardized service content. After all, the standard is equal to that of the public, and it is equal to that of the largest user.

AIGC's own characteristics actually limit the process of service standardization. Due to the instability of AIGC generated content and the uneven quality level of content, it is impossible to form a unified quality standard, which limits the expansion of user scale to some extent, and also limits the improvement of the profitability of AIGC enterprises.

As the AIGC track is in the exploratory stage, most of the technologies are not perfect enough to be used in production, and a small number of relatively mature applications are in the free trial stage to attract customers. This means that AIGC technology itself lacks liquidity.

The current capital is still too risky in the AIGC field. As a matter of fact, start-ups cannot break through the barriers of unstable AIGC technology, single application scenarios and lack of liquidity. In order to realize the rapid development of AIGC technology, the computing power and algorithm need to be greatly improved, which is difficult to predict.