At the Dartmouth Conference in the United States in 1956, four Turing Award winners including McCarthy, Minsky, etc., and a number of scholars jointly established; the concept of artificial intelligence is to hope that machines can recognize like humans. Know, think and learn, that is, using computers to simulate human intelligence. Since then, applications based on artificial intelligence have appeared. Since the 1970s, the emergence of artificial intelligence applications includes machine theorem proofs, machine translation, expert systems, games, pattern recognition, learning, robots and intelligent control, which are all imitating human intelligence. In this process, many schools have been formed, and what is now is; the connection school; the deep integration network. After the Chinese Academy of Engineering has studied artificial intelligence, especially in the application field, it has found that artificial intelligence is undergoing major changes. These major changes have brought about many new key theories and technologies, such as big data intelligence, swarm intelligence, cross-media intelligence, and human-machine hybridization. Enhanced intelligence, autonomous intelligence system. Big data intelligence is the artificial intelligence supported by the current big data; the Internet connects people to computers and people to form group intelligence; the cognitive method of cross-media processing has attracted more and more attention from the artificial intelligence community. ; The cross-media perception and computing generated by multimedia + sensors can be called cross-media intelligence; the combination of machines and people to form a more powerful intelligent system is called human-machine hybrid enhanced intelligence; liberated from the concept of robots, various Intelligent systems gradually developed into autonomous intelligent systems. In addition, there have been many changes in the application of artificial intelligence, such as smart manufacturing, smart cities, and smart medical care. How did these changes occur? We need to study its root causes. On this basis, the Chinese Academy of Engineering regards it as a major subject for research. The views put forward by this subject were accepted by the country, which led to my country’s release of the "New Generation Artificial Intelligence Development Plan" on July 20. Motivation for the replacement of artificial intelligence Why do we say that artificial intelligence is changing? We believe that the reason is that the world is undergoing tremendous changes: from the original physical space and human social space; the binary space has entered the ternary space with one more information space. How did the ternary space grow? Fifty years ago, the world was only a binary space, and all information flow and dissemination came from human beings. Even with the Internet, mobile communications, and search tools, it is still a binary space because the source of information is still people. However, today, a lot of information comes directly from the physical world-tens of thousands of satellites continuously convey information to the ground, hundreds of millions of cameras convey information through the screen, and a large number of sensors form a sensor network and become a new source of information. In the binary space, humans understand and transform the world through natural science and engineering technology; while in the ternary space with additional information space, humans can indirectly transform the physical world through human-computer interaction, big data, and automation of autonomous equipment. The ability is getting stronger and stronger. Meet the challenges and opportunities of AI 2.0 The changes in space have not only appeared big data, but also new channels. These new channels will bring new calculations and new social capabilities. This will not only provide new approaches and new methods for computer science and intelligence science but also form many new disciplines. For example, it is difficult for urban planners to plan the space, industry, and environment of a city in one go. However, understanding the city from the spatial level and through the channel of big data will definitely be able to understand more clearly how the city operates in a healthy way. Similarly, the complex environmental ecosystem, there are still many unknown medical and health systems, etc., are all; the complex system of scientific issues + engineering issues + social issues, it is difficult to understand them by traditional cognition and observation. Only by combining these methods with new cognitive methods can they be transformed into new ones. This is the basic motivation for artificial intelligence to move towards a new round of development. In terms of dynamics, the information environment has undergone tremendous changes, how can artificial intelligence remain unchanged? Artificial intelligence in the new information environment must be a new artificial intelligence. In terms of demand, human needs have also undergone tremendous changes. People need to use data methods to study smart cities to develop smart medical care, smart transportation, smart games, unmanned driving, and smart manufacturing. Artificial intelligence is needed to move from simulators to simulation systems. In terms of goals, from the pursuit of computer simulation of human intelligence in the past to the pursuit of human-computer integration, the pursuit of group intelligence that is more integrated with the Internet, humans and computers, this is the origin of artificial intelligence 2.0. I believe that with the expansion of information technology, there will definitely be new artificial intelligence technologies. What needs to be pointed out is that artificial intelligence moving from 1.0 to 2.0 is actually the deepening of the evolution of human living space from PH space to CPH space. There are many theoretical and practical challenges ahead of us. Technology that emerged in the AI 2.0 era Although artificial intelligence 2.0 is just the beginning, many new technical features have emerged. From the perspective of big data intelligence, deep learning technology is now very powerful, but it doesn't stop there. AlphaGo can cause worldwide shock not only because of its machine learning capabilities but also because of its use; new technologies such as self-game evolution, which is a new concept. It can be seen that in addition to deep learning, big data intelligence will also produce many new technologies. Some new applications are also very enlightening. A good example is that Google’s DeepMind team has been able to make money for Google; DeepMind uses its software to control 120 variables such as fans and refrigeration systems in Google’s data center. The software optimizes these 120 variables by reasoning. This has increased the power efficiency of Google’s data center by 15 and has saved Google hundreds of millions of dollars in electricity bills within a few years. In 2015, my country's data center consumed 100 billion kWh of electricity, which is equivalent to the annual power generation of the entire Three Gorges Hydropower Station, which is very enlightening to us. The use of computers to replace humans for organizational work has also emerged. What is not easy for one person or a group of people can be done by a group of wisdom. A project team from Princeton University in the United States has developed a game software called EyeWire. The goal is to color the connection between the human retina and the human brain through an electron microscope. However, this labeling of neurons cannot be done by a few people—there are so many nerves and every scientist only knows a small part of them. The project team used the Internet to call on eye neuroscientists from all over the world to jointly mark it. In the end, 165,000 scientists from 145 countries participated in this project. Humankind has unprecedented knowledge of the working mechanism of the optic nerve. This is the power of group intelligence. Human-machine integration technology-oriented hybrid intelligence also has great potential. Technologies such as wearable devices, semi-autonomous driving, and man-machine collaborative surgery have emerged on a large scale. This will become a new field and a large number of new products will appear. At the same time, cross-media reasoning has emerged. The cross-media technology of virtual/augmented reality has attracted attention in the past two years. Google glasses can; what you see is what you know, and instantly presents the information such as the origin and price of the objects you see; Microsoft's intelligent software can use photos to generate oil paintings and traditional Chinese paintings. This shows that cross-media technology is developing very fast. , Cross-media technology will greatly improve the intelligence of machines and people. In addition, unmanned systems are developing rapidly. In the past 60 years, artificial intelligence has vigorously developed robots, but robots/arms, drones, unmanned ships, etc. have developed rapidly. Many cities and companies have proposed replacing people with machines, but the core part is not to replace people but to make machines smarter and more autonomous. Therefore, autonomous intelligent systems still need a lot of research. The development of artificial intelligence 2.0 conforms to informatization; the development direction of digitalization-networking-intelligence. Many provinces, cities, and enterprises in China have formulated a new generation of AI development plans for their regions and units under the guidance of national planning, and are ready to do a lot of work. There is reason to believe that the rapid development of China's artificial intelligence technology and the industry is coming in unstoppable strides.