Big Data Digest welcomes all kinds of high-quality manuscripts

Please contact

Faced with the aggressive development of AI, the ethical issues of AI have become the focus of the times. This article focuses on emphasizing different perspectives of "Chinese and Western culture" and "science and humanities".

1. The ethical issues of AI have become the focus of the times

Today's AI has turned some topics that were considered sensational in the past into serious research that cannot be avoided. Not only are traditional scientific theory and technical practice and the relationship between the two being reconstructed, but also in the two major differences between "science" and "humanities". The challenging integration problem arises between the fields of AI. Faced with the aggressive development of AI, front-line scientists have to make expedient answers, led by the "Future Institute of Life" (of FLI) to formulate "Assi" The Lockheed Martial Artificial Intelligence Principles” ( ) can be regarded as an ethical emergency program for front-line scientists.

On the other hand, traditional philosophy, anthropology, ethics, etc. cannot quickly enter the forefront, and the frenzy stimulated by the industry, economy, finance, and social fields that are eager to be in trouble have risen to a national consciousness, and governments of various countries have We began to raise and take action on the development of AI and industrial creativity as strategic issues of global competition.

Unlike the research and development of AI in Western countries, Chinese culture attaches importance to the centrality of humanity, which allows us to have a special cultural position in the relativity of "Chinese and Western culture" and "science and humanity". Can we not lose the opportunity in the interweaving of such two "two cultures"? We would like to raise this question here. Faced with the human and machine ethical relationships that are closely related to human interests, we hope to study the ethical issues of AI platforms under a large pattern.

The development of AI and Robot has had an impact on the status and value of human beings themselves, and has been worrying about this for a long time. A little earlier, many funds and various research institutions have focused on AI ethics issues, including some famous artificial intelligence issues. The "Future Institute of Life" founded by experts and professors as volunteers published an open letter on January 11, 2015: for and: an Open letter: for robust and profitable artificial intelligence Research), proposing the slogan of "our artificial intelligence must do what we want them to do", that is, "helpful artificial intelligence" (AI). The same high concern for such issues has previously attracted widespread attention and serious discussions and actions from both inside and outside the industry, such as: IEEE for s in and (IEEE Global Initiative on Ethical Considerations of Artificial Intelligence and Autonomous Systems), AAAI 2008-09 on Long-Term AI (AAAI 2008-09 Presidential Committee on the Future of Artificial Intelligence), on to and (Artificial Intelligence Cooperation Benefits People and Society), AI100 (A Centennial Research Plan for Artificial Intelligence), of Fund (Artificial Intelligence Ethics and Governance Foundation), Allen for (Allen Institute of Artificial Intelligence), etc.

Earlier, before the emergence of physical robots, there were the famous "Three Rules of Robots" by science fiction novelist Isaac and many additional principles later:

LawⅠ:A Robot may not a human being or, , allow being to come to harm.

The first law: robots must not harm human individuals, or stand idly by witnessing human individuals that will suffer danger;

LawⅡ: A robot must obey given it by human would with the first law.

The second law: the robot must obey the command given by the person, with exceptions when the command conflicts with the first law; Law

Ⅲ: A robot must its own as long as does not with the first or law.

The third law: Robots should protect their survival as much as possible without violating the first and second laws.

These three laws actually regard machines as the principle of design and manufacturing of machines, and do not consider the problems of super artificial intelligence such as "autonomous ability" or even "self-awareness". Now, the "highly autonomous artificial" Intelligent systems (AI) or what is called "strong artificial intelligence" or "general artificial intelligence" (AGI) seem to be knocking on the door. How can humans deal with it?

In this sense, it can now be said that the cutting-edge and cutting-edge research conducted by any large company, large enterprise or academic organization for research and development of artificial intelligence are actually ethical issues related to artificial intelligence in essence.

2. The core idea of ​​"Asiloma's Artificial Intelligence Principles": AI

The "Asiloma Conference" (2017) held in California, USA in January 2017, including 844 experts in the fields of artificial intelligence and robots jointly signed what is called "Asiloma Artificial Intelligence Principles" (AI) , calling on the work in the field of artificial intelligence around the world to abide by these principles and jointly safeguard the future interests and security of mankind. The core idea of ​​the "Asiloma Artificial Intelligence Principles" is AI, which we interpret as "AI for humans and humans."

In the face of these complex and urgent situations, what humans can respond to is based on the basic ethical beliefs formed in human history and long-term ethical research results. Faced with the urgent ethical problems faced by the reality of the rapid development of artificial intelligence, a temporary answer from non-professional ethics scientists: "Asiloma Artificial Intelligence Principles", they are out of the traditional profession of scientists Morality, based on human conscience and rational concerns about the future of human beings, takes the ethical issues of researchers and institutions directly involved in artificial intelligence as the main object, and dispersed the ethical concept of AI in the "Asiloma Artificial Intelligence Principles" Among the 23 items, the programmatic entries are as follows:

The first "Research Purpose" determines the principle of AI research to create AI and Robot that serves people and is controlled by people. This principle is the basic ethical guarantee between humans and machines, and this guarantee is first studied and the ethical awareness of developers is reflected and abide by it in research and development; therefore, the 16th item "Human Control" expresses a deep expectation.

The second "Research Fund" conservation and protection of the premise of human resources, will and value systems.

Item 10 "Value Attribution", Item 11 "Human Values", and further improving the ownership of the value of machines to humans is the design principle of advanced artificial intelligence. Item 9 "Responsibility" will clarify this human-machine ethical relationship as a responsibility. The 12th item “Personal Privacy” and the 13th item “Freedom and Privacy” reflect human dignity.

Item 6 "Safety" is the conditions for the protection of machines to humans set for research work. Item 7 "Fault Transparency" and Item 8 "Judicial Transparency" are the existing research and development of artificial intelligence. Ethical requirements in terms of work system and legal system.

Item 14 "Share benefits", Item 15 "Common Prosperity", Item 17 "Non-subversion", Item 18 "Artificial Intelligence Arms Race", These are the manifestations of the ethical principles of human society in the field of artificial intelligence research.

The above contents are based on the human and machine ethical relationship between human and human nature, "Asiloma Artificial Intelligence Original

Then the last 23 articles of "common public interests" even include "super intelligence" (), but this is still considered as the overall interests of all mankind, which is artificial intelligence (AI) for the value and interests of human beings. The premise of maintaining human ethical status has been repeatedly stated at different levels in the field of artificial intelligence, all based on human morality and the simplicity of human dignity, but the development of machines or artificial intelligence is different from "nature" and What is the essence and ethical status of things different from human beings themselves is still an uncertain question.

The "Asiloma's Principle of Artificial Intelligence" is the awakening of the ancient human ethics spirit in the world's drastic changes, and an effort to defend human status and dignity in the world. The "Asiloma's Principle of Artificial Intelligence" is actually a concentrated expression of all long-term observations and thinking about the immeasurable impact of AI and Robot on humans, and in a more profound sense, it can be seen as A serious reflection on the nature, status and value of human beings represented by scientists.

3. Basic theoretical issues of ethics on AI platform

The most fundamental problem in the ethics of artificial intelligence is the ethical relationship between man and machine, and this ethical relationship between man and machine depends on the understanding of the meaning, value and status of "people". This is the oldest and most ancient Deep philosophical issues are actually the most fundamental problem in Eastern and Western philosophical thoughts by ethics. Philosophy disciplines and numerous knowledge disciplines, especially the development of science and technology since modern times, have caused philosophy and ethics to be constantly differentiated and dissipated into a wide range of disciplines. When modern people almost forget these ancient wisdoms, artificial The huge expectations, confusion and even fears faced by intelligence force us to push the ethical problems of artificial intelligence to the most important position.

In the most cutting-edge field of artificial intelligence research, the most confusing question is: Is there any "artificial intelligence" that is higher than human intelligence, the so-called "general artificial intelligence", "strong artificial intelligence", "advanced artificial intelligence" or " Super artificial intelligence, or even the new "superman" (Nicethells), "the animal that becomes a god" (Yuval Herali) is integrated with humans and super intelligence...? The 19th "Asiloma Principles of Artificial Intelligence" recognizes, but believes that this is an undeterminable problem. To attribute this problem to "uncertainty" is just a witty avoidance, that is, this The essence of the problem is: in the relationship between man and artificial intelligence, does this ethical relationship between man and machine based on basic human ethical beliefs? Therefore, everything seems to return to the oldest philosophical and ethical questions, what are our basic beliefs and knowledge of ethics?

So far, the entity construction of artificial intelligence is mainly based on the laws of physics. Physical entities do not have the problem of obeying ethical principles, but artificial intelligence based on physical entities has the intelligence of humans based on living entities. The essence is different, and this has always been the most difficult question in human knowledge theory, and people have been waiting for answers, but today, the ethical questions of artificial intelligence have forced scientists to make a temporary answer to deal with the artificial as a scientific and technological research activity. The basic ethical problem in intelligent work research and development. The "Asiloma Artificial Intelligence Principle" is just an emergency solution. Special theories such as philosophy, ethics, and anthropology must directly study the general nature of artificial intelligence, between people and machines. Ethical relationships, as well as the profound significance of human intelligence, emotions, and consciousness, elevate the technical relationship between people and machines to the theoretical height of ethics.

This is a serious challenge to philosophy and ethics. In a sense, this is the real battlefield for the challenges to people's status and future brought about by the rapid development of artificial intelligence.

The first philosophers in human history were also scientists. The ancient Greeks mixed the issues of science, philosophy, religion and ethics to discuss. After the Hellenistic era, global religions emerged, religions have the essence of ethics, and in the social reality in history. Ethical issues are mainly borne by religion, but religion does not answer human ethical issues, but only transfers the ultimate essence of human beings to God. Religion takes the transition from secular people to divine personality as its ultimate goal. Therefore, religion ultimately cannot replace ethics with the purpose of secular world.

The accumulation and progress of human knowledge and the development of science and technology have enabled modern people to enjoy unprecedented material "happiness", but knowledge and technology themselves have become a matter that is both objective and subjective. Pope once used The concept of "World 3" describes this particular object, but the theory of "World 3" has not received strong support and continued to be developed in depth. The difficulty is that the theory of "World 3" is just a conceptual system, not a system of traditional philosophy. In the framework, Pope did not provide a research model of tripartite relations that were different from the relationship between "master and guest" in traditional philosophy, so the relationship between people and knowledge, tools and technology has not been studied in depth and systematically.

Today, the development of artificial intelligence has covered and penetrated deeply into all things in human society, and has been directly exposed to the intelligence, emotions and consciousness that humans have always believed to be exclusive. The intelligent relationship between artificial intelligence and humans has become a value to people. and the challenge of meaning. The most cutting-edge research on artificial intelligence has become the theoretical frontier of ethics. Such problems are not faced and can be dealt with by individuals, collectives, institutions, enterprises or even individual countries. Obama accepts the joint efforts of Wired magazine. In the interview, it was pointed out that it is indispensable to play a role in AI development.

Today's researchers, groups, institutions, and enterprises have continuously adjusted, combined and reorganized different R&D platforms, and even national R&D platforms emerged, the ethical theory and practice of artificial intelligence have actually become the most important AI R&D platform on the AI ​​R&D platform An important public research topic realizes that building the ethical research of artificial intelligence as a cutting-edge theory of the platform is one of the most basic and basic tasks of platform construction.

4. Frontier ethical issues

Lao Tzu's "Tao Te Ching" and Aristotle's "Nicomaco Ethics" ( ) represent human deep thoughts about their own state of existence and social mode of survival. In this simple human consciousness, humans Always aware of the unconscious as the only self-conscious person in the world. Although "anthropocentrism" and especially the narrow "strong anthropocentrism" have been widely discussed or criticized, "man is everything." The simple concept of the unity of human values ​​and status has always been the dignity of man and man, but this ancient belief is facing crisis today.

If empirical science studies the world's "what" and "how it is", then ethics studies the "how it is" of society and what people should do. These two sciences have always been the two major issues of "science" and "humanity". The most representative discipline in the separate fields (the so-called "two cultures"), but the properties of computers, artificial intelligence and robots that emerged and developed rapidly from the field of scientific research have become unstoppable and have become human ethics. The erosion of the core concept of the machine "has" human intelligence, or the machine "has" human nature, does not seem to be a human fantasy. If the dignity and status of "human" are shaken by the development of human society itself, today it is also a bit However, it is in this ethical sense that the ethical issues of AI and Robot have become the focus of everyone's attention today.

The core concept of ethics is human social behavior, which is a moral philosophy in practice. The ethical concerns of AI and Robot actually have profound levels beyond this essence of human ethics:

(1) How can people (research, developer) implement traditional human morality and ethical beliefs in their research, manufacture, and develop AI and Robots, so that they can obey or even have human ethical principles;

(2) Do people have this kind of transcendent ability to do this?

(III) Are there or any ethical norms in AI and Robot over the physical laws that govern them? What's more, for example, can the so-called "science fiction model" super-artificial intelligence (Super-) ask such an ethical question?

In these complex and even unimaginable situations, we have to further ask: Are there any moral immoral or immoral issues of machines? Is this a human problem, or is it a problem between AI and Robot or a common problem of "we"? If you imagine this way, perhaps one day machines may have human-like or even their own nature different from human self, and even their thinkers and philosophers, who are human-like humans or vice versa, we are just humans. Subclasses in a class?

Before we transcend ourselves as the subject of this "human" or even "lifelike" we cannot and cannot conceive these problems in a substantial way. What is most urgent now is whether we have the moral mind to give AI and Robot the most moral mind. ability? Just as God could not have the nature of being exempted from sin when he created man, can we have this power that seems to be a power beyond God?

We can only explore on the cornerstone of the concept of human ethics that we already have, in this sense, we only have human ethics, and all our concerns, thinking and actions are human ethics. In this sense, AI is not the morality and ethics of machines, but just the ethical principles of human beings in researching and developing AI and Robot. Whether human beings can give human ethical principles to machines is an uncertain question. Today we can What we do is to adhere to the ethical bottom line of our human beings, so the "Asiloma Artificial Intelligence Principles" are human moral calls and the sense of crisis of human nature. In this sense, the "Asiloma Artificial Intelligence Principles" are the embodiment of human dignity. .

5. Research on artificial intelligence ethics from a global perspective

The "Asiloma Artificial Intelligence Principle" regards the ethical issues of participating researchers and institutions as the main project, while avoiding the ethical essence of human-machine relationships as an uncertain issue. On the one hand, it waits for artificial These essential problems in the development of intelligence are constantly exposed and gradually solved. On the other hand, we look forward to the modern revival of philosophy and ethics and the integrated development of theories about science, technology and the future of mankind and the world with a renewed vision.

The future of human uncertainty is always a lingering shadow, from the prophecy of Christian "apocalypse" to the Western irrational thoughts of "God is dead" (Nicethells) to "man is dead" (Foucault) and then to today The speculation of "the end of mankind" seems to be the sound of increasingly shaking footsteps. Can our culture and civilization be reborn in a new transcendent ethics?

Descartes said: "I think, so I am there." This is actually based on the scientific rational skepticism and critical spirit in modern times, which is a scientific attitude towards the world and people themselves, that is, the subjective world of one of the dualisms includes A subjective attitude of the objective world, including oneself, rather than a scientific explanation, and this subjective attitude does not equal the relationship between "I think" and "I am in", and other causes, such as cause and effect, logic, and rules, Even if there are countless research results in physics, physiology, psychology and other related aspects, it cannot explain how its own objectivity becomes the subjectivity of consciousness. Therefore, Western culture cannot be without a transcendent God. Apart from these, except for thought or consciousness itself, we cannot observe "thoughts" externally. We can only intuitively or realize "thoughts" and think "thoughts". Although we do not have similar difficult-to-understand quantum mechanics about unobservable "quantum states", Chinese philosophy or Chinese thought has always insisted that the reflection and transcendence of "thought" on the subject's inner (internal "mind") nature is the infinity of the intrinsic (internal "mind") nature of the subject The essence of the "thinking" of stepless (Wuji), only in this essence can we enter the path of understanding human "intelligence". Therefore, even if there is advanced or even higher-level artificial intelligence, it must be in the "behavior" of "thinking" that surpasses oneself, and the Chinese thought that cultivates Chinese culture itself has this kind of The inner temporal nature.

It is precisely because Chinese traditional culture has the essence of humanity and human nature that we can directly treat the ethical issues of artificial intelligence as essential ethical issues on the basis of inheriting traditional Chinese and Western theories and theories. Essential ethics Problems are the essential problems of people, and people are the essential uncertainty. The relative nature of uncertainty is certainty. Research on uncertainty issues can be carried out at different levels, from certainty to certainty. Only by the research and integration of various complex relationship patterns at the same level can we have hope of achieving real progress and progress.

References:

AI

()

on Long-Term AI

(

us/um///-2009.htm)

AAAI 2008-09 Panel on Long-Term AI

()

IEEE for in

and

()

for and :

an Open

()

on to and

()

Zhou Jianming, Liu Yu: Intelligent Philosophy: Ethical Issues of Artificial Intelligence (AI) Online Article

Zhou Jianming: Modern Ethics and Confucian Culture—Complete Chinese and Western Culture—Modern Article 1 Online Article