2024, China's AI application "big inventory"|Industry AI

Where exactly are AI applications traveling today? What are the general-purpose, tool-based, industry-based, and hardware-based? 

To this end, industrialists have done some inventory and combing of mainstream AI applications in China. I hope that while gaining insight into the cutting-edge trends of the industry, it will also provide an important window for grasping the future pattern of AI applications.

The data "surged to 910,000, up 264%, 34 times" comes from the statistical comparison of AI open source projects on GitHub.

For the development of AI application enthusiasm, the industry has never been so excited.

So, as of today, what are the domestic AI applications? In what directions and fields are they distributed? And their specific capabilities in the end how?

Statistically, AI applications can be divided into four categories according to application areas: general software, tool-based applications, industry software, and intelligent hardware. In the general software market, generative AI has taken the lead in office software, enterprise services, IT operation and maintenance, software development, network security, data intelligence and other applications, and has entered the pre-commercialization stage, with benchmark products appearing on the main tracks.

As the AI intelligent assistant (Coplilot) can deeply embed the ability of AI into specific application scenarios, it can actively understand the user's intention and provide molded solutions, becoming the most widely used product form of general generative AI application in China. Based on its innate advantages, the coworking field has more AI applications on the ground.

Tool-based AI applications mainly include chatbots, search engines, text tools, AI painting and code tools, etc., mainly focusing on the C-end.

It is worth noting that due to its high dependence on the underlying large model. The construction of competitive advantages mainly comes from differentiated product positioning and continuous training of more powerful underlying models and algorithms, so the degree of homogenization of domestic tool-based AI applications is currently high.

More powerful underlying models and algorithms are the key to building competitiveness for tool-based AI applications.

Industry software involves a number of industries such as finance, medical, education, industry, games, law, etc. Generative AI has more combinations in C-end scenarios such as games, law, education, e-commerce, etc., while the maturity of generative AI products in B-end scenarios such as healthcare, finance, industry, etc. is still low. At present, mainly financial, medical, education and other head manufacturers focus on building a large model of pendant class to promote the landing of related applications.

Intelligent hardware includes intelligent cars, robots, intelligent terminals, etc. At present, the combination of generative AI and intelligent hardware is mainly divided into two aspects.

One is voice assistant, the application scenarios include intelligent cockpit, intelligent speakers, household robots and other types of intelligent terminals. The other category is digital agent AI Agent, whose main applications include automatic driving, intelligent robots, etc., which has a broader application space. However, there are still bottlenecks in the perception and decision-making ability of AI Agent.

Overall, the overall development momentum of AI applications is still in the early stages. This can also be seen through the investment dynamics of the primary market.

According to CB Insights data, by the second quarter of 2023, the total investment in the field of generative AI compared to last year's $2.5 billion jumped 4.6 times, although about 70% of the investment funds are concentrated in the construction of the underlying AI infrastructure, including large-scale models, and the financing of the application layer accounts for only 30%.

In terms of the present, a question worth thinking about is, AI applications to where exactly? What are the general-purpose, tool-type, industry-type, and hardware-type? To this end, industrialists have made some inventory and combing of mainstream AI applications in China. I hope that in the insight of the industry's cutting-edge trends at the same time, but also to grasp the future pattern of AI applications to provide an important window.

Focus on the present, look to the future.