There will be 1000 AI unicorns in the hottest futu

  • Detail

Will there be 1000 AI unicorns in the future? Li Kaifu: impossible

Phoenix finance news, the 2019 China development high level forum was held at the Diaoyutai State Hotel in Beijing from March 23 to March 25. Elites from political, business and academic circles from all over the world gathered together to contribute world wisdom to China's development and help reform and development. Phoenix Finance reported the whole process on the spot, focusing on the summit dialogue of the world's top think tanks. Li Kaifu, chairman and CEO of Innovation workshop, attended the forum and delivered a speech

from the left of the above figure are: CEO of Phoenix, COO of Phoenix Satellite TV, Liu Shuang, chairman of Tata Group, Natarajan Chandrasekaran, CEO of Royal Philips of the Netherlands, Frans van Houten, CEO of byton automobile, Daniel kirchert, CEO of cheduoduo group, Yang Haoyong, chairman and CEO of innovation factory, Li Kaifu

Li Kaifu said that AI can be divided into four waves, They are Internet data, business big data, perceptual vision and hearing, and automated and autonomous artificial intelligence. Some venture capitalists get 10 to 100 times returns by investing in Unicorn enterprises

will there be 100 or even 1000 AI unicorns next? Li Kaifu said: the answer is absolutely No

Li Kaifu further elaborated that as a pure technology company, artificial intelligence is impossible to produce too many companies with a level of 10 billion dollars. Because as a technology, artificial intelligence has limited value in the transfer and sale of technology; Secondly, AI is a B2B technology, so its growth will not show explosive growth like some application products

the following is the transcript of the speech made by Kaifu Li, chairman and CEO of innovation works, at the 2019 China Development Forum:

I am very happy to share with you here. After publishing my new book, I have made 300 artificial intelligence speeches in the past six months. Today's topic is how AI reshapes society. I want to share these six questions

first of all, the first question is how much value AI has.

PwC and McKinsey have done some research and concluded that they can create a value of 100 trillion yuan in the next 11 years. These values use AI in every industry. In the Innovation workshop, we will divide AI into four waves, namely, Internet data, business big data, perceptual vision and hearing, and finally automated and autonomous artificial intelligence. You can see the industries listed below, which will really radiate to every field

of course, artificial intelligence is used in every field. Whether it is a disruptive use, reorganizing the whole industry, or whether it will bring a concept of injecting AI or ai+, this past and future will be different. If you pay attention to the field of artificial intelligence today, you mainly think about the unicorn of artificial intelligence. Those most valuable artificial intelligence companies, China has 17 unicorns of artificial intelligence, which we are very proud of. Innovation workshop invested in five of them, which brought us huge returns. Our investors got 10 times to 100 times returns in these investments

will there be 100 or 1000 AI unicorns below? The answer is absolutely No. Why? Because AI is a pure technology company, it is impossible to produce too many companies with tens of billions of dollars. Because as a technology, artificial intelligence has limited value in the transfer and sale of technology. Second, AI is a B2B technology, so its growth will not be like the explosive growth of melon seeds just mentioned by mark, because it is a consumer level use

where exactly did the 100 trillion yuan just mentioned come from? This is what we will see in the next stage of AI. At the beginning, AI is a black technology, which only top doctors can do. Then he expanded to do face recognition and so on, entered various vertical fields, and transformed these fields. The drooping fruit is almost taken. In the second stage, AI began to become a product, and it was also difficult to sell products, because each company had different needs. Now we are in this stage. We will see that the revenue growth of these companies will be slow. Next, AI is the same as Internet. At the beginning, Internet companies were of great value. We still remember the great value of these companies. But then slowly, every company is learning about interconnection. Finally, every company has an interconnection department, and finally, every department must have interconnection. AI is the same process, as AI gradually enters the mainstream from a very sophisticated and top-level black technology. In the mainstream process, the last and biggest beneficiary should be traditional industries. The example given by the CEO of Philips just now is a very good example. Artificial intelligence has been injected into a traditional company like Philips, and then created great value. It will enter different companies, so we believe that the inevitable way of AI in the future, that is, the way that the government should support, and the way that our investors pay more attention to, is how traditional companies use AI

of course, traditional companies use AI. A big problem is that traditional companies generally can't hire AI's top talents. Top talents go to Alibaba, Tencent, Google and other companies. What should we do? Let me share a secret with you. As long as you give me a great computer science graduate and give me five weeks, I can turn him into an AI engineer. We train 1000 talents every year, show you a hip-hop creation machine made by ten students who have never made products, give him a lyrics, give him a song title, and he can make such a song, and perform the cup making experiment of aluminum alloy plate, strip and foil with a thickness of 0.1 ~ 3.0mm. You can see Xiaobing's demonstration outside. We can do this with ten students without Microsoft Xiaobing's 100 top AI engineers. AI should enter the traditional, and traditional companies should embrace AI

third, what role can the government play in the growth of AI

first, it is certainly good for the government to support science and technology, but the government should pay more attention to the popularization of AI. Ai+ will become the most important trend in the future. Second, our AI talents, colleges and universities are generally difficult to cultivate relevant AI engineer talents. Using AI talents is very important, but many colleges and universities only want to do scientific research, and innovation workshops are willing to cultivate more applied AI talents. The Chinese government has made a very wise decision, technological pragmatism, which is a point mentioned in my book. Don't worry too much about AI. European friends manage AI and slow down. China's approach is to promote AI first, and then deal with it when problems are found. This is a very wise approach

the last point is infrastructure. Just now we talked about byton's cars. If the roads in Jiangsu or the cities in xiong'an can have the function of strengthening driverless safety, even if China's technology is not as good as the United States, it can go on the road faster, collect data faster, iterate faster, and even become the best driverless in the world

fourth, AI brings such great value

we have a huge poverty alleviation. In the future, the expansion speed of market capacity will be accelerated and popularized. But AI will bring many problems. Here are many problems. I won't talk about them in detail today. Pick two examples. The most common example is privacy. When you think about privacy, you think of gdpr made by the European Union. Gdpr is an incomplete design, which slows down AI and allows users to choose which functions of which stations I am willing to authorize. Users can't do such tedious things. Considering personal privacy, it is not a privacy issue, but a personal data management issue. First of all, we should manage the data well. According to the policy of the Chinese government, the transfer and sale of data will be severely punished. This will prevent Facebook from happening in China. We let users choose the right to protect data, but we can't let users choose each function of each station. We must use technology to overcome technical problems. The picture on the right is a picture of my personal imagination. Just like our ability to turn the voice up and down, we should also combine the protection of personal data with the relative convenience we obtain. Behind it is AI to help you protect your data. In the era of network security, what should we do every time hackers attack? Is it the way of governance to catch hackers? No, this does not include the disturbance of viruses and crashes. We must use technology to prevent them from entering

Innovation workshop is developing federated learning, which is a technology that can protect your privacy and still achieve the goal of machine learning. Another issue that needs more attention than privacy is security. Security issues, face recognition is very powerful, but wearing glasses and a few stickers on your face can fool face recognition. What should I do? On the left is a video generated by the Obama administration. In fact, he didn't say it, but it seems to be true. What about these? These technologies need to be differentiated by higher technologies of artificial intelligence. It is impossible to rely on human faces or legal methods alone. The most worrying thing is that AI continues to lower the threshold. After the threshold is lowered, how to avoid potential disasters in the future? We must reach an international consensus, and use technology to overcome technology and standardization. These innovative workshops are involved, including my own co chairman of the Artificial Intelligence Council of the world economic forum, which is also for this

the fifth problem is artificial intelligence. Some companies say that it will bring a lot of unemployment, and all jobs are gone. Some companies say no, artificial intelligence will only help mankind. The answer is all wrong. We consider creativity and caring. Caring is the communication between people. High and low are our y-axis, and the degree of creativity is to create. We can see that in the upper right corner is the human world, in the lower right corner is the opportunity of man-machine integration, and in the lower left corner, low caring and low creative work will be completely replaced. You should worry, but don't worry too much

what should we do? I am the first of all hydraulic universal experimental machines and electronic universal experimental machines. We should give the people in the lower left corner a chance to find another chance to find a new job. Many people say that every technological revolution brings more opportunities, but it's not that simple. This time, repetitive work will be replaced, while AI creates non repetitive work. In addition, we should also consider that there are opportunities for caring work. For example, in the United States, two million caring jobs will be produced in the next five years, such as nurses, elderly care, and home care. These company jobs can only be done by people. Why don't you do this job? Because the income is too low, people with a sense of society and the government should change these conditions

finally, is there a universal model of AI governance? Here I would like to say to the foreign guests here, in fact, what the West particularly cares about is individualism, which is a very valuable and respectable value. But in fact, many countries in the East are collectivist countries, and we will never have a way to govern the world. Everyone is hidden

Copyright © 2011 JIN SHI