Artificial Intelligence: Embracing Change and Protecting Value

  [Kehai Tide Watching]

  In the past few years, a large number of innovative technologies have emerged in the field of artificial intelligence, and key breakthroughs have been made in the fields of autonomous driving, voice and image recognition, medical diagnosis, intellectual competitive sports, etc. Intelligent scenes that people once could only imagine in science fiction movies are quietly coming, and an intelligent era full of infinite possibilities has been fully opened.

  Artificial intelligence with several ups and downs

  Artificial intelligence is not a new concept in science. As early as half a century ago, computer science giants represented by Turing, Weiner, McCarthy, Shannon and so on began to explore artificial intelligence scientifically. Turing took the lead in putting forward the basic proposition of artificial intelligence "Can machines think" in his paper "Computing Machines and Intelligence" published in 1950, and designed the famous "Turing Test" — — If a machine can start a dialogue with human beings without being identified as a machine, then the machine is intelligent, which has become the first research to define the standard of artificial intelligence from the scientific level. At the same time, Wiener put forward the famous "cybernetics" and discussed the unity of machine and human — — People or machines accomplish a certain purpose through feedback, which reveals the possibility of simulating people with machines and lays an important foundation for the proposal of artificial intelligence. In 1955, McCarthy, Minsky, Shannon and others, as co-sponsors, invited a group of scholars to discuss Artificial Intelligence at Dartmouth College in the United States, and used the term artificial intelligence, which led to the rise of artificial intelligence scientific research.

  Although artificial intelligence has always attracted the attention of the scientific community, due to the constraints of basic technical conditions such as computing power, algorithms and data, artificial intelligence technology and commercial development have not been smooth sailing, and even once fell into a trough. Until May, 1997, IBM computer system "Deep Blue" defeated Kasparov, the world champion of chess, which once again triggered global attention to artificial intelligence. Since then, Professor Hilton of the University of Toronto in Canada has made a key breakthrough in the field of artificial intelligence with deep learning as the main technical line. Artificial intelligence has been applied in natural language processing, speech recognition, image recognition, machine translation, fraud detection, product recommendation, etc. In addition, mobile Internet, cloud computing, big data and Internet of Things have matured and formed a technical community effect, providing a technical environment for the technological breakthrough of artificial intelligence. A number of IT giants such as Google, Microsoft, IBM, Facebook, Baidu, Tencent, etc. also foresaw the commercial dawn of artificial intelligence and accelerated the industrial layout of artificial intelligence. In October 2016, the then US President Barack Obama released the report "National Strategic Plan for Artificial Intelligence Research and Development" at the White House Frontier Summit. This strategic plan, known as the new "Apollo Moon Landing Plan" in the United States, marked the US government’s emphasis on artificial intelligence, and artificial intelligence has reached a turning point of breakthrough.

  According to the development stage and level of artificial intelligence, some foreign experts divide artificial intelligence into three stages: Artificial Narrow Intelligence (ANI), Artificial General Intelligence (AGI) and Artificial Super Intelligence (ASI). Weak artificial intelligence is good at unilateral ability, such as only having the intelligence to play chess. Strong artificial intelligence refers to artificial intelligence that has the same cognitive ability as human beings in all aspects; Super artificial intelligence refers to surpassing the limits of human intelligence in all fields. Generally speaking, human artificial intelligence is still in the stage of weak artificial intelligence moving towards strong artificial intelligence.

  Cutting-edge technology that subverts the future

  As a subversive technology with "singularity" characteristics, artificial intelligence can not only drive a new round of scientific and technological innovation and economic growth, but also overcome a series of development difficulties faced by human beings in the fields of transportation, environment and health care, thus creating a new paradigm for global economic and social development.

  In 2016, Accenture, a famous consulting firm, released a report saying that artificial intelligence will become the driving force for a new round of global economic growth. By 2035, by changing job attributes and creating new man-machine relationship, artificial intelligence will double the economic growth rate, and the problems that have plagued the world economy for a long time, such as production stagnation, resource mismatch and labor shortage, will be solved, and people will be able to get out of "inefficient" work and do truly efficient and creative work.

  Intelligent transportation and unmanned driving are the key areas of artificial intelligence development at present. In the future, the transportation facilities with artificial intelligence application will not only include passenger cars, but also trucks, aircraft and personal robots. Automated transportation will be everywhere, and cars will become better drivers than humans. Intelligent transportation systems will effectively overcome urban traffic jams and parking difficulties, reduce the incidence of car accidents and reduce automobile emissions, and there will be fewer private cars for urban residents, thus changing the overall appearance of the city and bringing a brand-new form of urban organization.

  Artificial intelligence will also make great achievements in the medical field, which will help mankind overcome a series of medical problems to a great extent and solve the problem of uneven distribution of medical resources. Take IBM Watson as an example. An expert from the Medical College of the University of North Carolina in the United States asked Watson to analyze and diagnose more than 1,000 cancer cases. It was found that Watson gave the same treatment plan as a human oncologist in more than 99% of the cases. A Tokyo woman with leukemia was sentenced to death by a doctor, but Watson read the relevant 4,000-meter-thick medical materials within 10 minutes and gave a treatment plan to save her. Recently, a teenager in Britain underwent medical surgery based on artificial intelligence technology and got rid of epilepsy for seven years. The doctor asked the robot to drill into his brain, stimulated the cerebral cortex through electrodes, and found out the part that caused the seizure. The doctor removed the lesion.

  An out-of-control problem of no trial and error

  When artificial intelligence technology continues to reshape people’s work, study and lifestyle, and brings infinite happiness to mankind, it is also generating more and more open problems.

  1. Artificial intelligence is out of control. It is an ancient and ultimate problem of artificial intelligence whether it can keep control of an artificial intelligence system that is far smarter than human beings one day. The problem of control and out-of-control is so important because human beings may not have the opportunity to try and make mistakes at all. Although it is generally believed that the existing artificial intelligence technology is still in the stage of weak artificial intelligence, many experts believe that the era of super artificial intelligence is not far away, and the form of super artificial intelligence can realize recursive self-improvement and lead to the final "intelligent explosion". In 2008, a survey of experts at the Global Catastrophic Risk Conference held in Oxford University showed that machine super-intelligence has surpassed nuclear war, various manipulated epidemics and climate change to become the greatest threat to mankind. By 2100, the possibility of this artificial intelligence leading to human extinction is 5%. In this regard, the author of "A Brief History of the Future" and Israeli historian Herali pointed out three possible directions for human beings: human beings will lose their economic and military uses, and the economic system will no longer be valuable; For the system, human beings as a whole will have value but individuals will have no value; However, the system will find value in some individuals, but they will become a new species of superman, not the general public. "System" refers to a new social form produced by the rapid development of biological science and information technology.

  2. Machine discrimination. In the Internet of Everything environment, artificial intelligence systems based on massive data and powerful algorithms are influencing or even replacing individual behavior decisions. Under the trend of artificial intelligence replacing individual decisions, algorithm deviation and machine discrimination are not uncommon and have great influence. For example, Google’s photo software has wrongly labeled black people as "gorillas", and Flickr’s automatic tagging system has also wrongly labeled black people’s photos as "apes" or "animals". The problem of Google’s algorithmic discrimination has attracted people’s attention. In Google search, it is easier to find advertisements implying a criminal history than searching for white people’s names. In Google’s advertising service, men see more high-paying job advertisements than women.

  3. Skilled unemployment. Technical unemployment is also a major challenge arising from the development of artificial intelligence. Kaplan, a professor of artificial intelligence and ethics at Stanford University, believes that the most important challenge for human beings to deal with artificial intelligence lies in the change of labor force and the distribution of wealth in the whole society, which has an impact on the stability of the existing society. Schwab, chairman of the World Economic Forum, also pointed out that the wide application of high technology in computer-controlled, fully automated and networked production and service industries, such as the use of robots, may lead to the loss of a large number of jobs in the future.

  4. Human-machine emotional ethics. The development of artificial intelligence technology will lead to the subversion of traditional human ethics. In the autumn of 2015, the British BBC and CNBC reported that an American company had created a female image "sex robot", and the person in charge of the company believed that the product could help the widowed solve their life troubles. It is reported that the robot presupposes a variety of personality patterns, such as "communicative" and "obedient", but the product has aroused extensive ethical discussion.

  As Hassabis, the founder of DeepMind, said: "Artificial intelligence is a great technology. It is neutral in itself and may be used in good places or bad places, so we must ensure that its users are responsible." In this regard, the United States, Britain, Japan, as well as the United Nations, the European Union, IEEE and other international organizations have issued strategies, laws, standards and norms of artificial intelligence, hoping to promote the healthy development of artificial intelligence technology and industry through the guidance of artificial intelligence policies and ethics. We should scientifically regulate artificial intelligence from laws and regulations, social security, technical standards and application norms, so as to promote the healthy development of artificial intelligence technology and industry and prepare for the future of artificial intelligence.

   (Author: Hui Zhibin, deputy director of Internet Research Center of Shanghai Academy of Social Sciences)