For the development of AI, Li Kaifu, chairman of the Innovation Workshop, has been a well-known optimist.
In an exclusive interview with Davos, he firmly believes that although AI will replace some of human work, there are four areas where human values ​​are irreplaceable: need to be smart, need to be creative, need to be cared, and need Relatively complex strategic cross-domain thinking.
However, he also warned that when humans use AI to create the greatest wealth in history, they will face the same big challenges. "If we don't do anything, then in 15 years we will inevitably encounter the greatest challenge in human history, even the biggest disaster." The solution he gave was to re-train the replaced human beings to be valuable. Work, while companies, VCs, and even governments should also make a difference in providing more jobs.
In Davos, the contrast between China and Europe and the United States on AI has become a topic that has to be talked about.
Li Kaifu told reporters that the United States is undoubtedly leading the world in AI. Although Europe has advanced technology, due to restrictions in the entrepreneurial environment and government policies, it is bound to lag behind China and the United States in the development of AI. China has its own advantages in capital, market, data, and policies, and will have the opportunity to catch up with the United States.
The following is a record of reporter dialogue Li Kaifu:
European thinking or restricting AI development, China has the opportunity to catch up with the United States
Reporter: Thank you for accepting an interview with Sina Finance and Sina Technology. You came to the annual meeting of Davos to participate in a lot of discussions about artificial intelligence, and there should be a lot of sparks in the middle.
Kai-fu Lee: Yes, this time there is a lot of communication with the founder of DeepMind and Google's Li Feifei. Google is undoubtedly leading the world in AI. I think that Google, a company that may lead the world, should not say that China and the United States are strong.
AlphaGo zero talked with DeepMind, they are very frank, although this is a game that can surpass humans for thousands of years in the game without training data, and can change the rules, but they also think that this machine learning is not immediately applicable A lot of scenes, it is only a game that is relatively more accurate and fairer than Go, and a standard reward system. Other fields may be able to do it, but it may not be possible in three or five years.
When talking with Li Feifei, she said that the dream of the future is to see the combination of man and machine and human-machine collaboration. I personally think it is quite difficult, because from the perspective of capital and business, we must want to use AI to replace humans or use it to make more money. To do this kind of man-machine combined with 1+1=3 work, more people need to do it, but the school can't afford it, the big company doesn't think it makes money, and the capital won't enter. I think Li Feifei also mentioned a very good question, that is, who is going to implement this dream and vision. Are we too short-sighted in artificial intelligence?
Reporter: Li Feifei said in a sentence that artificial intelligence may now be a good dish, but it may burn the kitchen. Does this mean that it takes a long time for artificial intelligence to reach our ideal state?
Kai-fu Lee: This time in Davos, we also discussed several problems caused by artificial intelligence, including security, privacy, inequality, replacement of work, and prejudice. Many people have discussed so many challenges and immature Under the premise, how should we move forward? This topic actually has a standard Chinese school debate, and there is a standard European and American faction debate.
The Chinese party’s debate will say that security issues must of course be noted, but in addition to security issues, the wave of technology is coming too fast, entrepreneurship must be implemented and promoted quickly, and AI requires a lot of data, so let’s first Things will be made and then the data will be better and better, we can adjust the error at the beginning.
The thinking in Europe and the United States will say that we can't do it with such a big thing. We must debate from all aspects. How can we ensure that an unmanned driver is safe from the moral level? For example, one person in front is sure to be hit. At that time, there are two people who may be hit, and the AI ​​itself is very difficult to make this judgment. European and American thinking will hope that every different voice can be heard, there is a rational, deep, logical thinking debate, and finally reach a consensus, and then use this consensus to advance our technology. So these two ways of thinking are very significant differences.
Reporter: The difference between Chinese thinking and European and American thinking in artificial intelligence will lead to differences in policy and development speed between the two sides?
Kai-fu Lee: It is not just a policy. I think entrepreneurs, investors' thinking, teachers, professors' thinking, and of course government policies are all inside. Based on the European and American way of thinking, policy orientation must first discuss the problem of inequality between the rich and the poor. It is necessary to consider that the privacy data should be in the hands of the individual, and each data must be authorized. They will want to give back data rights and then rewrite business rules. If everything is over, this model may be the most effective way, but today Google and Facebook have made so much money in Europe, and then the user data is in their hands. You ask him to spit it out. This is How can it be done?
Reporter: There are rules against the business world.
Kai-fu Lee: Yes, so I think this debate is particularly interesting. Of course, we also see that the EU can do a fine to force Google and Facebook to commit crimes. What kind of challenges will this bring to their business model, and what opportunities will it bring to Chinese companies? Because Google and Facebook do not He will make a system for Europe, and he may be global. For example, Tencent can make it easier to enter Africa, because Facebook has made this thing very complicated, so I think this will bring a lot of variables.
In addition, I think that as a citizen of the world, it is not right to criticize this European-style thinking, because Europe has such a long history, such a good system, such a good culture, etc., also comes from this strong consciousness. . But today's technology era is running faster than at any time in history. Will this kind of thinking make him fall behind? I think this is also a very interesting observation.
Reporter: Will this kind of thinking bring some obstacles to the development of European AI?
Kai-fu Lee: Europe has its own advantages, but it also faces many challenges. In terms of advantages, the education of European AI is actually very good. There are excellent professors in Cambridge and Oxford. They have made good results, but they are all partial. In terms of entrepreneurial environment, London is already the top in Europe, but VC is less, VC returns are low, VC is conservative, and because of the language cutting, Europe is not a big market, so entrepreneurship in the UK Just think of the UK as a market. For example, I talked to the British VC. They would like to see 100,000 pounds of income before the A round. I asked him what is the significance of the 100,000 pounds income? He said that we have to look at the British market. We don't have data on the amount of data and activity. We only have to look at the income data, so I think they are pushing themselves to a dead end, a small market, a pattern of views. Not enough, VC money is too small, the return is not high, although there are good technology, but there seems to be no way to reverse the situation that it must lag behind China and the United States.
Reporter: Compared with Europe, China has very abundant capital and huge data. The government also supports AI. Does China have an opportunity to catch up with the United States?
Kai-fu Lee: I think China and the United States should have similar opportunities. Of course, today the United States is leading, and China will successfully catch up, but there are still many variables to surpass.
Judging from my wave of four-wave artificial intelligence, in fact, some of China are leading and some are backward. The first wave is the Internet AI, which China used to be backward, but now it has surpassed the United States. We can see that in the benchmark valuation of the giants, in addition to Baidu, Google, China has begun to be equal to the United States, and should be exceeded in the future. In the second wave, it is to use existing commercial data to create commercial value in the fields of banking, insurance, medical care, etc. In this respect, China is backward and will continue to lag behind because China’s data is chaotic and unorganized. To mobilize it to do a good job in banking, insurance or hospitals, there is a gap between the United States and the United States. The third wave, I call it the digital world, that is, using video and voice to capture data and sensors that I have not had in the past, using the data to make smart decisions, such as using cameras and sensors in unmanned stores. Know who is coming, who bought it, who saw it, who didn't buy it, and then use this data to make the offline store as smart as online, or it can be used in classrooms, intelligent transportation, hospitals, clinics, etc. field. But this kind of business all over the camera, I think Europe and the United States is very difficult to accept, so China will lead in this regard.
The last is automated AI, which is the application of the unmanned, robotic industry. This United States still has an advantage now, and China may catch up, but it is not necessarily. The threshold of technology is the highest, and China has a bigger gap to catch up with. There are other advantages in the United States. For example, to make money and save money for commercial robots, the wages of workers in the United States are relatively high. The income may be five or seven times that of China. It is easier for American robots to replace people in the United States. Because the cost does not need to be too low, and in China, the cost will be lowered again, so it may land first in the United States.
Of course, we know that Shenzhen is a "world factory". It can go from being humanized to being underhumanized and unmanned. It can also be seen in China that there is a better policy push for driverless driving, so that testing can happen. For example, a truck driver in the United States protested that when the car was unmanned, don't measure the truck. Otherwise, we are all unemployed. Trump's policy promised, because the truck driver is mostly Trump's voter, he can't. The 3 million voters are completely ignored, but the probability of at least occurring in China is relatively low.
Reporter: The government's executive power will be stronger.
Kai-fu Lee: Yes, the probability of this happening will be relatively low. These are the advantages of China. So I think it’s really hard to say. China and the United States should have opportunities. If we continue to develop these advantages, we just talked about it. The advantages of data, policies, and capital, plus we still need to carefully consider these security and privacy issues, because even if we make a lot of companies to make a lot of money, as long as there is a serious crisis, such as 100,000 users The money was stolen, or the driver was hijacked by the bad guys, and then became a murder weapon. Once this happens, we may go back three or five years. Therefore, I still feel that although we have the opportunity, the concerns of these Europe and the United States are real. We should use the methods in line with China's national conditions to see how to avoid these disasters.
Reporter: As far as the development of Chinese AI is concerned, BAT has a huge advantage. So, are there any opportunities for entrepreneurs to break through?
Kai-fu Lee: I think the opportunity is still a lot, because AI covers all industries, many industries BAT is not concerned or has no advantage, such as medical industry, driverless, industrial AI, is not good at BAT. So from the revolution of the entire AI, BAT should have the best position to get the most benefits, but because the field is too big, many entrepreneurs will still find some advantages that BAT does not see or can not leverage, or this field The chances of making a profit are too small, because BAT has to take into account the return and opportunity costs, so BAT may not necessarily put people on this emerging business. But for a startup, if you have a good income in the segment and then grow well, a VC will be willing to invest a lot of money in the company, so I think the opportunity is still some.
Four areas of AI are difficult to replace people, but the failure to replace people with AI will bring disaster
Reporter: For the development of AI, some people still have some fears, such as AI replacement, then where is the person's position and how should it survive in society?
Kai-fu Lee: For example, as a truck driver.
Reporter: Right.
Kai-fu Lee: I think this is a very good question, because today AI is gradually replacing human tasks, and then it will replace some people's work. The challenge to human beings in this process is not only to say that income is reduced, but also The challenge of self-existence, respect, and self-fulfillment.
The value of the truck driver lies in the ability to drive the goods to where it is on time. This makes him think that he is contributing to human beings and is valuable. Suddenly, the machine is better than the people. What should he do? So I think we have to consider retraining the existing careers that may be affected, and we may have to consider social welfare. Social welfare is not to say that blind money is given to everyone, but that if you are willing to learn, then we can look for some industries that will not be replaced for you to do. I think there are four areas that AI is hard to replace. One needs to be smart. One needs to be creative. One needs to be cared for. The other is that it requires relatively complex strategic cross-domain thinking, so from children. In terms of education, these four must quickly explore its advantages. From a training perspective, it may be difficult to train a particularly creative scientist, or a strategist or CEO, but to train a craft is flexible, as long as it is more flexible than a machine, such as a plumber, Or aircraft aerospace workers or repair experts, these positions should not be replaced by AI in the short term.
Or if we consider the care and love industry just mentioned, whether it is going to a nursing home, or an orphanage, or a hospital to do some work such as taking care, rehabilitation, or even volunteering, or many traditional industries, such as tour guides and masseurs. In the service industry, I think that these jobs are actually economically valuable on the one hand, and trainable on the other. Thirdly, I think it is very important that they are basically helpful people. Helping people will bring us a sense of satisfaction. It may even exceed today's truck drivers. For example, a truck driver he received retraining, he went to the nursing home to help the elderly, the latter's satisfaction is the same or more intense.
Reporter: But the development of AI is very fast. You mentioned that these vocational trainings, including our training for the next generation, will take time. In the process of this time difference, there will be a hostility between humans and AI. Is this a period of anxiety?
Kai-fu Lee: I think one thing is to see if everyone can accept these new jobs and then be retrained. This is definitely a difficult process. Because people are inert, people are told that your work is worthless. You have to Finding a new job and training you will also bring about some problems like depression. I think this may be the biggest problem. We must do our best to do it.
As a government agency or a company, you can also do something, such as Shuangchuang. In the past, it was all about creating a technology company and then creating the next BAT. But in fact, I don’t think every entrepreneur can do BAT. Can we have a new social enterprise creation and VC, they are dedicated to investing in and supporting these companies that can provide employment opportunities, that is to say, 100 jobs will be provided this year, 500 will be available next year, and 1,000 will be available in the following year. If there are such a group of companies entering the direction of our double creation, receiving various government subsidies, or a group of new VCs are loving VCs, they are not purely for return, but to create value for society, I think these Everything is something we need to think early. Although I am an optimist myself, I have described many visions for the future, but I have clearly seen it. If we do nothing and do nothing, then 15 years later we should face the greatest human being ever. The challenge is even the biggest disaster.
Reporter: This year's annual meeting of Davos is entitled "To create a common destiny in a divided world." In fact, this common destiny is not only between the countries, but also between humans and AI.
Kai-fu Lee: I think you are very right. This topic is the best Davos theme, because no matter how the country divides or even how to compete, we all face the same challenge. When we create with AI, When human beings have the greatest wealth in history, we are faced with these same problems, and its solutions are actually diversified.
We can see that, for example, a craftsman may be highly respected in Switzerland and Japan. Some service industries are also highly valued in some countries. For example, volunteers and volunteers in Taiwan are doing very well. We can also refer to, for example, Myanmar and other Southeast Asian Buddhist countries. He has different methods to satisfy people. In fact, it is a bit of a struggle between the world. Even everyone who works two or three hours a day is very happy. Finally, Everyone is happy. Each of our models can actually be referenced. So I think that Davos should bring more wisdom around the world to discuss how to face this problem, enhance the overall wisdom of mankind, and then find some solutions that can be used all over the world.
Plug-In Connecting Terminals,Insulated Spade Terminals,Cable Connector Double Spade Terminals,Vinyl-Insulated Locking Spade Terminals
Taixing Longyi Terminals Co.,Ltd. , https://www.lycopperlugs.com