On the afternoon of May 20, the 23rd Shanghai Social Science Popularization Activity Week·Roundtable Forum, hosted by the Shanghai Ethics Society and co-organized by the Law School of East China Normal University and the Ethics and Wisdom Research Center of the Department of Philosophy of East China Normal University, was held at East China Normal University. The theme of this forum is "Ethics and Legal Principles in Artificial Intelligence Governance". More than 100 experts, scholars, teachers and students from Fudan University, Shanghai Jiao Tong University, East China Normal University, East China University of Political Science and Law, Shanghai Normal University, Shanghai Academy of Social Sciences and other institutions participated in the discussion, and more than 5,000 viewers participated in the forum online.

The opening ceremony of the forum was hosted by Fu Changzhen, professor of the Department of Philosophy of East China Normal University and vice president and secretary-general of the Shanghai Ethics Society. She pointed out that a new generation of artificial intelligence is booming around the world and is profoundly changing human production and lifestyle. While the transformation of artificial intelligence technology improves human well-being, it also poses a series of new challenges to human society. In response to abnormal events and emergency ethical problems that arise in artificial intelligence governance, there is an urgent need for in-depth joint responses across disciplines and fields.

Professor Zhang Zhiming, Dean of the Law School of East China Normal University, delivered a speech at the opening ceremony. He pointed out that artificial intelligence is developing rapidly and has promoted the upgrading of various fields of society as a new productive force and new business format. He emphasized that the complexity of artificial intelligence governance requires interdisciplinary collaboration to deal with, and the combination of philosophy and law can make governance issues more in-depth. Existing research mostly regards artificial intelligence as the object of legal relations, but strong artificial intelligence may require a new legal subject framework. He specifically pointed out that artificial intelligence legal research cannot be just lively but not in-depth, as this will only lead to academic bad money driving out good money. Simple anthropocentrism cannot solve the regulatory dilemma, and human security and technological development need to be balanced.

Wang Guoyu, Distinguished Professor of Fudan University, Dean of the Institute of Science and Technology Ethics and the Future of Humanity, and Director of the National Applied Ethics Professional Degree Graduate Education Steering Committee, said in his speech that the relationship between ethics and legal principles is an enduring theme in the history of philosophy. In today's era of artificial intelligence, this issue has become more urgent and more realistic. Ethical, legal and social issues need to be put into the same framework for in-depth discussion.

Gao Guoxi, President of the Shanghai Ethics Society and Distinguished Professor of Fudan University, delivered a speech on behalf of the organizer, the Shanghai Ethics Society. He said that ethics and law are both ways of social adjustment and are two different forms of value pursuit that dominates a society. The problem-centered joint research between law and ethics is of great significance. This in-depth exchange and cooperation between the Law School and the Ethics Society is also an important measure to serve the national strategy and promote knowledge innovation.

The rapid development of artificial intelligence brings many challenges

The keynote report session was hosted by Professor Hu Yuhong, professor of the School of Law of East China University of Political Science and Law, dean of the Institute of Human Rights, and editor-in-chief of "Law". Three well-known scholars from different fields shared their research results.

Ji Weidong, director of the China Institute of Law and Society at Shanghai Jiao Tong University, said that since November 2022, the rapid development of generative artificial intelligence has brought about many problems. AI’s international resources and knowledge expand the space for legal choices, but also bring new ethical and legal risks. Ji Weidong suggested that in the era of large models, AI governance should adopt multi-level and diversified legal norms to balance security and development to ensure that the goals of the AI ​​system are consistent with human values.

Cheng Sumei, researcher and deputy director of the Institute of Philosophy of the Shanghai Academy of Social Sciences, discussed the ethical challenges brought by algorithmic systems in a report on the theme of "The Dilemma of Responsibility in Artificial Intelligence Algorithm Systems." Cheng Sumei pointed out that the nonlinearity and complexity of algorithmic systems, as well as the open source characteristics, make traditional causal tracing and responsibility attribution mechanisms face challenges. "Dark knowledge" in complex algorithm systems not only aggravates the information cocoon effect, but may also make algorithm errors difficult to foresee. The pervasive use of algorithmic systems may obscure persuasive intentions and put people at a disadvantage. Facing the impact of AI, it is necessary to transform from an instrumentalist to an interactionist view of technology and create a responsible ethics.

Researcher Duan Weiwen, director of the Philosophy of Science and Technology Research Office of the Institute of Philosophy, Chinese Academy of Social Sciences, pointed out that the current understanding of AI ethical risks is often speculative and metaphorical, which blurs our true perception of AI risks and may also lead to AI ethics discussions falling into "ethical bleaching", "ethical hype" and "ethical idling". The current characterization of AI ethical risks is mainly based on verbal expressions such as "privacy" and "discrimination." In the semantic space, the fixed connection between these concepts and vocabulary limits our understanding of the dynamics of AI risks. As a result, he used the mutation theory/catastrophe theory model to propose a new perspective on "cognition-intervention" of AI risks.

Rebuilding a new ethical relationship of human-machine symbiosis

The first session of the special report was hosted by Professor Yu Feng, associate professor and deputy director of the Department of Philosophy of East China Normal University. Four well-known scholars from different fields shared their research results.

Yang Qingfeng, a professor at the Institute of Science and Technology Ethics and the Future of Humanity at Fudan University, said that although strong artificial intelligence was once considered a topic in the field of science fiction, technological progress has greatly increased the reality of this topic. Yang Qingfeng outlined three approaches to achieve super intelligence: from general intelligence to super intelligence, from embodied intelligence to super intelligence, and from interactive intelligence to super intelligence. Facing the challenges that superintelligence may bring, he analyzed the alignment method as a technical response and pointed out the possible misunderstandings of value and technology in this method. As an ethical response, he proposed a "contract ethics", believing that in the future we and superintelligence will be in an ethical relationship of mutual shaping and mutual understanding in coexistence.

Fu Changzhen, professor of the Department of Philosophy at East China Normal University and executive deputy editor-in-chief of the Journal of East China Normal University (Philosophy Edition), said that with the advent of the digital intelligence era, human civilization is facing unprecedented changes. We need to rebuild a new ethical relationship of human-machine symbiosis by "studying the interface between man and machine", re-understand the nature of human beings, let AI continue to develop in the direction of "adapting human values" and achieve value alignment, pursue the unified theory and cognitive architecture of artificial intelligence with practical wisdom, and create general intelligence (AGI) with independent perception, cognition, decision-making, learning, execution and social collaboration capabilities, and in line with human emotions, ethics and moral concepts.

Wang Jun, associate professor at East China Normal University School of Law, said that the application of artificial intelligence in administrative law is still in the stage of policy experimentation and normative preparation. The widespread application of artificial intelligence, such as automated administrative approval and full-process automatic penalty mechanisms, not only improves the efficiency of law enforcement, but also brings new legal challenges. The current discussion can be carried out from the perspective of "small incision, big care" and a method based on dogmatics. In the future "legislative moment of artificial intelligence governance", we should pay attention to both intelligence and legalization, balanced development and security, and ensure the protection of the legitimate rights and interests of citizens, legal persons and other organizations when using artificial intelligence technology.

Chen Zhaoxin, associate professor of the School of Law at East China Normal University, pointed out that the rise of artificial intelligence has brought new challenges to legal professional ethics and legal practice. He mentioned the concept of legal enhancement, which uses artificial intelligence technology to help lawyers overcome cognitive and emotional biases and optimize the legal decision-making process. He also believes that artificial intelligence can not only assist legal predictions, but also play an important role in improving legal understanding and decision-making levels, but attention should be paid to maintaining the normativeness and independence of the law.

Explore issues in the areas of ethics and law

The second session of the special report was hosted by Hu Yan, deputy editor and deputy director of the Journal of East China Normal University. Four experts and scholars jointly discussed the cutting-edge issues of artificial intelligence in the fields of ethics and law.

Yan Hongxiu, a professor at Shanghai Jiao Tong University's Digital Future and Value Research Center, discussed "Is value alignment the 'only way' for humans to lead to the future?" He believed that the development of artificial intelligence requires value alignment, but phenomena such as illusions, jailbreaks, and deceptive alignment often occur in the value alignment process. "Alignment" is best understood as the process of "alliance" or "mutual calibration," and the alliance between the two is actually our expectation for humans to dominate in the human-machine relationship.

When Duan Lei, associate professor and deputy dean of the Law School of East China Normal University, discussed "Legal Dilemmas of Using Artificial Intelligence in the Investment Field," he emphasized the practicality of artificial intelligence in law. He also shared the legal challenges encountered in using artificial intelligence in law using commercial law as an example. Duan Lei introduced an example of a company using artificial intelligence tools to make investment decisions. In the era of weak artificial intelligence, artificial intelligence is only an auxiliary tool for natural person directors to make business decisions. As long as there are no prohibitions in the articles of association, natural person directors can authorize some decision-making matters to artificial intelligence for decision-making. Even when necessary, in order to maximize the company's interests, natural person directors also have the obligation to let artificial intelligence make decisions.

Cui Zhongliang, associate professor at Nanjing University of Information Science and Technology and postdoctoral fellow at East China Normal University, discussed "Social Risks and Governance of Generative Artificial Intelligence as a Narrative Subject" and pointed out that the autonomy and complexity of artificial intelligence, as well as people's acceptance and trust of artificial intelligence subjects as they grow in human-computer interaction, are proving that generative artificial intelligence has the potential to become a narrative subject. This phenomenon will bring risks at the individual, organizational and societal levels. In this regard, he put forward three governance suggestions: ensuring human subjectivity, limiting the application of artificial intelligence in specific fields, and cultivating self-reflection capabilities.

When discussing "Is artificial intelligence a taxable subject?" Li Shuai, associate professor and deputy dean of the School of Law at East China Normal University, discussed the EU's 2016 recommendations and pointed out that AI's replacement of human labor may lead to tax reductions and challenges to the social security system. Through specific cases, he demonstrated the impact of AI on corporate taxation and analyzed the possibility of taxing AI. He suggested treating AI use as a taxable activity or directly taxing its income, which requires further exploration at the legal and technical levels.

How to establish human uniqueness

The meeting entered the review phase, where five experts made in-depth comments. Xiao Yang, a professor at Kenyon College in the United States and a visiting professor at East China Normal University, focused on the issue of giving moral or legal status to artificial intelligence and robots. Cai Zhen, a professor at the Department of Philosophy of East China Normal University, discussed how to establish human uniqueness in the era of artificial intelligence. She believes that the concept of "rationality" in moral philosophy has the connotation of self-legislation, and human will also has a reflective structure and is much richer than the concept of rationality at the artificial intelligence level currently being discussed. Pan Bin, a professor at the Department of Philosophy at East China Normal University, believed in the review that there is an important logical approach between artificial intelligence from ethical initiatives to legal regulations, that is, how to make artificial intelligence more explainable. From the perspective of the legislative necessity of artificial intelligence, the transparency and explainability of AI are the prerequisites for legislation. However, the complexity of AI itself, the necessity of intellectual property protection, and professionalism limit the transparency of algorithms and the explainability of models. Professor Li Ping, deputy director of the Economic Theory Research Center of Henan University of Finance and Economics, pointed out that with the rapid development and widespread application of artificial intelligence technology, allowing artificial intelligence to understand human values ​​​​and achieve real-time two-way human-machine value alignment has become a research area that has attracted much attention. Today, many research teams and experts have proposed aligning AI values ​​​​with human values. However, there are many problems in this process. How to choose applicable ethical principles according to specific scenarios is an important issue that requires continued in-depth research. Wang Taoyang, associate professor of the Department of Philosophy at East China Normal University, believed in the review that the arrival of strong artificial intelligence poses two challenges to ethics and law. First, the stage of strong artificial intelligence poses challenges to ethics and law itself. Second, humans have extremely high requirements for artificial intelligence.

Pictures provided by the school