【Guangming Youth Forum】

Editor's note

On February 21, 2023, the Chinese Ministry of Foreign Affairs officially released the "Global Security Initiative Concept Document", calling for "strengthening international security governance in emerging science and technology fields such as artificial intelligence, and preventing and controlling potential security risks." In the process of Chinese-style modernization, technological innovation of artificial intelligence is one of the important forces to promote my country's scientific and technological innovation. As the most representative disruptive technology, while artificial intelligence brings potential huge development dividends to human society, its uncertainty will also bring many global challenges and arouse major ethical concerns. General Secretary Xi Jinping pays close attention to the development of emerging technologies such as artificial intelligence, and emphasizes the need to accelerate the improvement of "governance capabilities in fields such as artificial intelligence security" and "shape the cultural concept of science and technology for goodness, so that science and technology can better enhance human welfare." To this end, this edition specially organized several young scholars to discuss the ethical challenges and scientific response of artificial intelligence, and invited experts to comment, in order to arouse more attention from the academic community and contribute wisdom to promoting the healthy development of artificial intelligence.

Talk to the person

Peng Jiafeng PhD student at the School of Philosophy of Renmin University of China

Yu Hao PhD student in the School of Politics and International Relations, East China Normal University

Deng Yulong PhD student in the Department of Philosophy, Nanjing Normal University

host

Liu Yongmou Professor of the School of Philosophy of Renmin University of China and Researcher of the National Development and Strategy Institute

1. Artificial intelligence that coexists with opportunities and challenges

Host: The new technological revolution is in full swing, and emerging technologies represented by artificial intelligence and other technologies are developing rapidly, greatly expanding the scope of time, space and people's cognition. Human beings are entering an era of intelligent interconnection between "human, machine and matter". Please tell me in detail what opportunities artificial intelligence brings to the development of human society?

Peng Jiafeng: Intelligent technologies such as artificial intelligence, big data, Internet of Things, and cloud computing are booming, which has had a profound impact on all aspects of human society and has promoted the entire society to gradually enter an intelligent society. In this process, there are many major historical opportunities that we need to grasp. As for technological governance, artificial intelligence, as a governance technology, is promoting changes in the governance concept, governance methods, governance efficiency and other aspects of social governance, elevating traditional technology governance to a new stage of intelligence, presenting the "intelligent governance" Comprehensive" trend. Intelligent governance will comprehensively improve the intelligence level of social public governance, and mainly present four characteristics: First, governance integration, that is, promoting the integration of various intelligent technologies and other governance technologies, and greatly improving the governance level of intelligent society; The second is governance dataization, that is, based on the growing massive amount of data, the governance goals are achieved by social computing on the "digital world" mapped by data; the third is governance precision, that is, the powerful perception and transmission of intelligent technology Capability and computing power transform traditional extensive governance into precise governance; fourth, governance algorithmization, that is, continuously improving the intelligent decision-making system, and trying to expand programmatic algorithmic decision-making into more decision-making activities, thereby improving the quality of decision-making.

Yu Hao: Artificial intelligence helps reflect on the foundation for the establishment and development of human society. As analytical AI continues to evolve towards generative AI, especially generative AI has initially shown functions such as discriminating problems, analyzing emotions, developing dialogues, and creating content. The fields that originally belonged to humans are being employed by artificial intelligence. It gradually erodes with another set of computer languages ​​composed of "0" and "1". This is not only an impact on human society, but also will inevitably enhance human subjectivity in a more equal and open framework and promote the further development of human society.

Deng Yulong: Overall, the development of new technologies represented by artificial intelligence has significantly improved social productivity. For example, generative AI can not only complete the analysis and judgment of traditional AI, but also further learn and complete creative work that analytical AI cannot do. From the perspective of human-computer interaction, artificial intelligence has also promoted the efficient development of production relations. Specifically manifested in: First, it stimulates the transformation of labor forms. Artificial intelligence efficiently undertakes a large amount of basic mechanical labor, while human labor transforms into higher-level creative labor, which leads to the transformation and upgrading of the labor structure at the social level, and uses artificial intelligence as an intermediary to integrate and coordinate labor within the society. Upgrades are also achieved. The second is to promote the reconstruction of the labor field. With the transformation of labor forms and the expansion of socialization of labor, artificial intelligence frees labor from fixed fields, and the flexibility of human labor increases. Compared with creative labor, mechanical labor is more restricted by space and time. After artificial intelligence replaces basic labor with lower marginal costs from the technical level, the freedom of human labor space and time has achieved a leap. Third, higher requirements are put forward for the development of the subject, especially for the subject to adapt to social development. The development of artificial intelligence technology challenges the traditional knowledge structure of human beings, requiring human beings to update the original knowledge structure to adapt to the needs of social development, and also put forward higher requirements for education. The educational model and educational content need to be more in line with the level of scientific and technological development and cultivate More comprehensive talent.

Host: While a series of products of artificial intelligence bring convenience to people’s lives, they also arouse people’s vigilance against the ethical challenges they may arouse to a certain extent. Some people are concerned about the risks of artificial intelligence and are somewhat anxious about the advancement of artificial intelligence. How do you view this vigilance and anxiety?

Yu Hao: The risks of artificial intelligence and the anxiety caused by it are completely understandable. But we cannot return to a world without artificial intelligence. Artificial intelligence has already deeply intervened in human society. Trying to curb the advancement of artificial intelligence can only be a mantis arm to get the car. At the same time, we cannot let it go about the development of artificial intelligence. Ignoring or even suppressing the advancement of artificial intelligence can only cover up the truth. Therefore, we should face up to this anxiety, explore solutions in the process of developing artificial intelligence, and seek opportunities in danger in the risks brought by artificial intelligence.

Deng Yulong: We should view this anxiety correctly. We should see that anxiety has its positive significance. It reflects human sense of crisis, gives rise to foresight thinking about the risks of artificial intelligence, and reminds us to pay attention to the problems existing in the development of artificial intelligence technology behind anxiety. Correct treatment of anxiety can help actively take measures to prevent risks, dialectically analyze prescient thinking in anxiety, and resolve risk problems through the upgrading of social governance models. At the same time, anxiety and fear alone are not enough. What is more important is to actively solve the social problems brought about by the development of artificial intelligence. From the perspective of labor, artificial intelligence will indeed replace some human labor, promote the transformation and upgrading of the labor structure, and let labor develop towards fragmentation and individualization. Workers are in a weak position and face the challenge of "replacement of machines with people". But we should also rationally realize that artificial intelligence is not a complete replacement for human labor capacity, but puts forward higher requirements for workers, requiring workers to master scientific knowledge, internalize the development of technology into their own abilities, and in the future, Realize your own value in creative labor.

Peng Jiafeng: The invention and use of any technology will inevitably be accompanied by one or another risks. Artificial intelligence technology is naturally no exception. In its application process, it also triggers risk issues such as privacy leakage, algorithmic discrimination, and legal liability. Therefore, paying attention to the risk issues of artificial intelligence and thus creating anxiety about the advancement of artificial intelligence has a certain theoretical basis and practical basis. But we should be more aware that some related risks of artificial intelligence can be avoided in advance and do not necessarily occur; even if they do, effective means to resolve risks can still be continuously sought. Taking the risk of personal privacy abuse as an example, although it is inevitable to involve the collection and analysis of personal data during the governance process, personal privacy can be protected and the risk of abuse can be reduced by establishing a complete normative and regulatory system.

2. The "ethical track" of competition in artificial intelligence technology

Host: General Secretary Xi Jinping pointed out when attending the 15th G20 Leaders' Summit via video, "China supports strengthening dialogue around artificial intelligence, and proposes to hold special meetings in a timely manner to promote the implementation of the G20 AI principles. Lead the healthy development of global artificial intelligence.” Please talk about what the "Artificial Intelligence Principles" should include? What practical value does the cultural concept of science and technology have in promoting the healthy development of global artificial intelligence?

Peng Jiafeng: In order to cope with the ethical challenges brought by the rapid development of new technologies such as artificial intelligence, in 2022, the General Office of the CPC Central Committee and the General Office of the State Council issued the "Opinions on Strengthening the Ethical Governance of Science and Technology", which clarified "enhancing human welfare" and "respect Five scientific and technological ethical principles, including the right to life, "adhering to fairness and justice", "reasonable risk control", and "maintaining openness and transparency". I believe that these five principles basically cover the ethical requirements of the principle of artificial intelligence and demonstrate the cultural concept of science and technology for good. The fundamental goal of the cultural concept of science and technology for good is to enable science and technology to better serve the society and the people and bring good social or social welfare. Technology for goodness has at least the following three practical value in promoting the healthy development of global artificial intelligence: First, it is to shape public trust. The public's trust in artificial intelligence is largely not entirely determined by the degree of relevant risks, but depends on whether the public's interests and values ​​are paid enough attention. The latter is the inherent requirement of technology for goodness. The second is to lead technological innovation. The cultural concept of science and technology for good will play a value-leading role in the process of technological innovation and development. The third is to promote global cooperation. The cultural concept of science and technology for goodness attempts to establish the "greatest common divisor" of artificial intelligence ethical norms on the basis of reaching ethical consensus, countries can establish mutual trust and achieve more full and in-depth international cooperation.

Yu Hao: Personally, I believe that the principles of artificial intelligence should also include the concepts of non-confrontation and non-out of control. Non-confrontation means that artificial intelligence should not be regarded as an adversarial existence in human society. Artificial intelligence has become a constitutive element of human society. We must face artificial intelligence with a more open attitude. Non-out of control means that ethical norms of artificial intelligence should not be given up, and artificial intelligence that is accelerated development should be standardized in an intelligent way. If the above concept is taken as the premise, that is, in the case of supporting the development of artificial intelligence, the cultural concept of science and technology for good becomes extremely important in promoting the healthy development of global artificial intelligence. The "good" here points to "good governance" at the national governance level. When the development of artificial intelligence expands from the national scope to the global scope, "good governance" has a more realistic connotation in the sense of building a community with a shared future for mankind. All countries should abandon the Cold War mentality and zero-sum game, and think together based on goodwill and friendship, how human beings as a whole can lead to global "good governance" under the impact of artificial intelligence.

Deng Yulong: In 2019, the EU issued the "Trustable Artificial Intelligence Ethical Guidelines", and in 2021, China's National New Generation Artificial Intelligence Governance Professional Committee issued the "New Generation Artificial Intelligence Ethical Code" (hereinafter referred to as the "Normative Code"). Compared with the ethical norms issued by the EU, the Norms reflect the institutional advantages of socialism with Chinese characteristics and aims to integrate ethical norms into the entire life cycle of artificial intelligence. The fundamental purpose of the development of artificial intelligence is to promote the all-round development of people. Therefore, I believe that the principles of artificial intelligence should also reflect the requirements of shared and orderly development. Sharing, aims to prevent the technology monopoly of artificial intelligence. Science and technology development should take into account the interests of all people, rather than serve minorities, and all people share the achievements of scientific and technological development to promote the common growth of global science and technology levels. Orderly development aims to prevent the disorderly expansion of artificial intelligence technology. The development of artificial intelligence technology is ultimately to enhance people's sense of happiness. Promoting the orderly development of science and technology can promote the harmonious integration of human-machine machines and effectively prevent the risk of potential disorderly expansion.

Host: From the normative perspective, what are the main aspects of ethical reflection in regulating the development of artificial intelligence?

Peng Jiafeng: In recent years, major countries around the world have increasingly fierce competition in the field of artificial intelligence, and have placed the development of artificial intelligence at the strategic level of national development. For example, the United States has successively issued the National Artificial Intelligence Research and Development Strategic Plan (2016) and the Executive Order on Maintaining the United States' Leadership in the Field of Artificial Intelligence (2019); the European Union has successively issued the European Artificial Intelligence Strategy (2018) and "Artificial Intelligence White Paper" (2020); China also released the "Three-Year Action Implementation Plan for the "Internet " Artificial Intelligence" (2016) and the "New Generation Artificial Intelligence Development Plan" (2017). An objective situation of competition in artificial intelligence technology has been formed. In this context, if the global risks and challenges brought about by the development of artificial intelligence technology are ignored, it is very likely to fall into the competitive logic of technological catch-up. Therefore, it is urgent to regulate the technological competition of artificial intelligence, and advocating ethical reflection may be a feasible path. The significance of ethical reflection lies at least in: First, set the ethical bottom line. The development and application of artificial intelligence technology need to follow some basic value concepts and behavioral norms. Only by keeping the bottom line of ethics can it be possible to avoid the occurrence of subversive risks. The second is to achieve agile governance. Ethical reflection is a dynamic and continuous process that runs through the entire life cycle of artificial intelligence technology activities. In order to ensure that it always serves the original intention of improving human well-being and technology for good, it is necessary to maintain the due moral sensitivity, and resolve the various ethical challenges brought by artificial intelligence in a flexible, timely and effective way, and ensure that it is good for good in technology for good. The road to move forward steadily and achieves healthy development.

Deng Yulong: The competition in artificial intelligence technology is to promote the development of science and technology, and the ultimate goal of science and technology development is to promote the progress of human society. The competition in artificial intelligence technology should not only include a single dimension of technological competition, nor should it curb the development of science and technology in other countries through technological advantages. Instead, it should be a comprehensive competition under the conditions of artificial intelligence technology, and promote global artificial intelligence and all mankind through healthy competition. common development. This should include social governance competition, ensuring social fairness through social governance, and thus ethical reflection on the relationship between people in society constitutes an organic part of the competition for artificial intelligence technology. First of all, ethical reflection puts higher requirements on the competition for artificial intelligence technology. The fairness, trustworthiness, interpretability, transparency and security of artificial intelligence are not only ethical requirements, but also represent the development direction of artificial intelligence technology and are the technological commanding heights that need to be seized by the competition for artificial intelligence technology. The development of science and technology is for the comprehensive development of people, so human development is embedded in the requirements of scientific and technological development, and ethical reflection helps to prevent the proliferation of instrumentalism. Secondly, ethical reflection provides value guidance for the competition in artificial intelligence technology. Ethical reflection focuses on protecting people's rights. Scientific and technological development is not the only measure of social development. We should also pay attention to the diversity factors, especially the protection of the interests of special groups, such as preventing adverse effects such as data gaps. Ethical reflection helps to achieve the comprehensive and healthy development of artificial intelligence.

3. Artificial Intelligence Security and the Comprehensive Development of Humans

Host: Scientific inquiry has always been an important cognitive way for people to understand the world and understand themselves. How does the revolution in information industries such as artificial intelligence affect people's cognitive ways?

Peng Jiafeng: The revolution in information industries such as artificial intelligence has promoted the birth of a new paradigm of scientific research - data science, and has thus had a profound impact on people's cognitive methods. Data science is considered to be a new scientific research paradigm after experiments, theory and simulations. Compared with traditional science, data science combines statistical and computational thinking. Through massive data, powerful algorithms and computing power provided by technologies such as artificial intelligence, it can directly find correlations, extract correlations or predictive knowledge from the data, and then generate A scientific mindset based on relevance. But this correlation does not necessarily translate into causality, because interpretability is crucial to extract causal explanations from correlations determined by data science technology, and related technologies generally lack the necessary transparency and interpretability. Data science is more likely to become a prediction science, but prediction is not the only goal pursued by science. Explain and intervene in phenomena by revealing the potential causal structure of the world are also two important goals of science. Therefore, although data science can generate relevance knowledge by analyzing large amounts of data, it cannot directly generate causal explanations. In this regard, traditional science's testability hypothesis methods and exploration of causal laws still have important value. Data science does not replace traditional science. On the contrary, the two will complement each other and become an effective tool for human beings to explore the world.

Yu Hao: It is obvious that as artificial intelligence moves towards general artificial intelligence, the educational resources, life entertainment and work information it can provide to people is becoming more and more abundant, and people will inevitably rely more and more on interacting with artificial intelligence. To obtain external information. Therefore, when artificial intelligence deeply forms a filter for people to understand the world, if they do not remain alert to the cognitive framework that has repetitive and homogeneous tendencies in artificial intelligence itself, artificial intelligence may distort people's cognitive ways until it affects people. Subject creativity.

Deng Yulong: The new technological development represented by artificial intelligence is called the Fourth Industrial Revolution. The most prominent feature is the deep integration of machines and humans. Machines are no longer used by humans as an external tool. It is the way of cognition that influences humans in deep association with humans. On the one hand, the information industry revolution has enriched the connection methods of human cognition. The development of artificial intelligence and big data technology promotes the expansion of human analytical logic from causality to correlation, and the emphasis on correlation allows artificial intelligence to obtain information from big data rather than small data dimensions, providing a new perspective for human cognition. According to the traditional way of human cognition, causality requires that the cognition of the world is certain, which is difficult to achieve in the complex society of the digital age. Artificial intelligence's perception of correlation fills this gap, allowing us to predict future trends without being able to grasp certain information but mastering a large amount of data. On the other hand, if we blindly trust the output results and generated content of technologies such as artificial intelligence, and make an absolutely equivalent connection between the results and content and empirical facts, and mistakenly believe that it is the whole of facts, then we will lose the humanistic abstraction We should be vigilant about the ability to reflect and always adhere to the humanistic spirit of reflection and criticism.

Host: How to adjust the relationship between people’s subject creativity and information sharing?

Peng Jiafeng: When people gradually hand over more creative work to artificial intelligence, it is inevitable that people will worry about whether artificial intelligence will threaten human subject creativity. From the perspective of human-computer relationship, this concern is based on a human-computer hostility perspective, believing that artificial intelligence squeezes the creative space of people's subject and is a continuation of alternative logic. But from the perspective of human-computer collaboration, we regard artificial intelligence as a powerful helper for humans. By creatively using artificial intelligence, we can give humans greater creative space. For example, when doing text writing, multimedia scripting, program code, document translation, etc., artificial intelligence can first complete the draft work at a high level, and then humans can make some creative adjustments and exercises. At this time, the content generated by artificial intelligence will become the raw material for further creation, and humans will invest in creative activities with higher efficiency. Of course, it is not easy to achieve the above effects. It not only requires a change of ideas, but also requires corresponding adjustments in institutional arrangements, educational methods, etc.

Yu Hao: Faced with artificial intelligence with highly integrated and shared information, humans may become animals of algorithms. Imagine the following scenario: When activating based on artificial intelligence becomes convenient and effective enough, actors will tend to accept artificial intelligence. At this time, it seems that human actors are interacting based on natural language, but in fact, algorithm logic is based on computers. Language is operating digitally. As a result, the creativity of human subjects is eroded, and humans may become algorithmic animals. We should remain sober and vigilant enough for such situations.

Deng Yulong: Content generated by artificial intelligence technology (AIGC) has the characteristics of highly integrated and shared, and can efficiently mine data and generate information on human knowledge. We need to adjust the relationship between people's subject creativity and information sharing in a high degree of integration and sharing in information. First, we need to expand the AIGC database through human creativity. At present, AIGC mainly relies on large language models and uses a large number of network texts. As a training database generation, creative generation through humans can be further expanded to improve its richness by further expanding the training text of the database. Secondly, it is necessary to provide value training for AIGC through human creativity, and integrate the value stance, ethical laws, etc. generated by human creativity with the AIGC training database to build a trustworthy and sustainable information highly integrated sharing mechanism. Finally, it is necessary to use the content generated by human creativity to be used as the source of human knowledge with AIGC. The acquisition of human knowledge cannot be limited to AIGC alone, but requires people to use their subject creativity to reflect and expand the content generated by artificial intelligence technology, and to transform humans. Undata-implemented, empirical knowledge and AIGC-data-implemented knowledge become the source of human knowledge.

(Compiled by Zhang Yingtian, editor of this edition)

"Guangming Daily" (15th edition, April 10, 2023)