Advantages and disadvantages of the hottest AI sup

Advantages and disadvantages of AI "super power"

in recent years, AI has made significant progress in many fields, such as image, voice, natural language processing, computer vision, driverless, robot, and even go and games. These developments have made AI related enterprises and the capital keen to invest in AI see the value and urgency of promoting the development of AI technology. As a result, the profit seeking nature of capital and the dependence of technology on capital have led to a significant increase in the capital injection of AI by existing technology companies in this upsurge, and the rapid growth of Unicorn companies is also closely related to massive capital investment. Moreover, with the support of funds, a large number of outstanding scientific researchers who were originally engaged in basic research related to artificial intelligence in the academic community have also been attracted to the industry. It can be seen that super powers supported by technology giants, excellent researchers and huge amounts of capital, such as foreign Google, Microsoft, domestic bat and a few Unicorn companies, continue to dominate in the field of artificial intelligence, and even expect to dominate

on the positive side, the formation of super power has promoted application level technological innovation and made people feel the rapid progress of artificial intelligence. In recent years, the rapid popularization of various face recognition systems in China's high-speed railways and airports, the significant improvement of the automation of parking fee collection systems, and the omni-directional substitution of payment for traditional financial transactions are all significant advances in artificial intelligence technology. p>

in terms of commonness, these advances come from the maturity of AI in related technologies, or at most, there are only a few technical difficulties, but they do not affect the formation of practical products, such as security related products. In the formation of derivatives, we do not need to make fundamental changes, and we can quickly form products, such as applications related to face recognition. Therefore, it will be faster and more convenient to promote the development of these AI applications through super power. Moreover, the formation of super power can also ensure sufficient financial resources to develop some expensive frontier fields, such as driverless, smart cities, etc. At the same time, the concentration of advantageous talents is also conducive to tackling difficult problems that need to be solved by large teams, such as the development of large aircraft. These are undoubtedly the good things that super power can bring through the experiment

however, everything has its pros and cons. If there is a positive impact, there must be a negative side. From the perspective of super power itself, out of the need for company operation and responsibility for capital, super power tends to choose research that attracts attention and can help the company quickly increase value or improve valuation. For this reason, in terms of technological development, super power prefers research or application that is certain, low-risk and easy to see results in the short term. From the perspective of sense of gain, super power will also tend to protect some mature new technologies and methods that may affect the market sales of existing products, which will lead to the lag of sense of gain

from the perspective of ecological environment, it will also affect the orderly development of small enterprises. This influence comes from two aspects. On the one hand, it is led by the super power to evaluate the value of emerging enterprises and carry out subsequent financing according to the rules of the game formulated by the super power. For example, in the field of artificial intelligence, capital may be evaluated according to the number of papers published by enterprises in the field of artificial intelligence. On the other hand, small enterprises are squeezed in the space of technology and scientific research. For example, the research results formed by small enterprises, which spend a large proportion of enterprise funds, may soon be copied by super power with 90 degree and 180 degree stripping power through talent advantage. In addition, it lacks the market share owned by super power. For plastic materials, unless it is necessary or new materials, as a result, the R & D results invested by small enterprises cannot recover costs or profits from the market. Therefore, Small businesses are likely to be reluctant to spend too much money to support such research or applications. This indirectly affects the all-round and diverse development of artificial intelligence

from the perspective of artificial intelligence, there are still many problems that have not been effectively solved, some of which are not even clear about the reasonable technical route or theory. For example, the imitation and learning of common sense intelligence are still lack of fundamental breakthroughs in the understanding of insight, inspiration and brain activity mechanism, which are the core concerns of our customers. Artificial intelligence and even the emotional research of pet dogs are still at a loss. These are not what super power is willing to spend time and money to touch

it is worth mentioning that on February 14, 2019, nature published a research paper on scientific and technological breakthroughs and team size by Wu of the University of Chicago and Northwestern University. After analyzing more than 65million papers, patents and software during 1954 and 2014, they found that large teams tend to form incremental achievements in mature frontier fields, while small teams and independent scientists tend to form disruptive innovations in immature frontier fields that can form new ideas and opportunities. This conclusion applies not only to academic research, but also to industry to a certain extent

therefore, in order to achieve the balance between super power, small enterprises and schools in the research and application of artificial intelligence and form a variety of output results, on the one hand, we can rely more on super power to continue to work on large and relatively mature technologies, so that the public can have more sense of gain, happiness and security, on the other hand, we should also consider giving more preferential policies to small enterprises, Encourage them to engage in research and application in artificial intelligence or related fields to ensure their good survival and development space. In addition, the school will be given long-term scientific research funding support, such as the National Natural Science Foundation of China, which is of great significance to the school's scientific researchers and can provide stable scientific research funding support, to ensure that they can freely explore in the relatively immature frontier fields that require 5 to 10 years or even longer to see success

[the author of this article is a professor and doctoral director of the school of computer science and technology of Fudan University]