VOL.3451

From April 25th to 26th, 2024, the International Academic Conference on the Ethics of Artificial Intelligence (AI: A Cross-) was successfully held at Fudan University.

Carl Mitchum (Carl), the first president of the International Society for the Philosophy of Technology and a professor at the Colorado School of Mines, President of the 18th International Society for the Philosophy of Technology, a member of the EU High-Level Expert Group on Artificial Intelligence and a professor at the University of Vienna, Mark Kaukolberg (Mark), a professor at the Technical University of Darmstadt in Germany ( ), and Armin Grunwald (), chairman of the Technology Assessment Bureau of the German Parliament and a professor at the University of Karlsruhe ) and Christopher Kernan, editor-in-chief of the international journal Nano and Emerging Technology Ethics and researcher at the Institute for Technology Assessment and System Analysis (ITAS) at the University of Karlsruhe ( ), Professor Gong Ke, former president of Nankai University and executive director of the China New Generation Artificial Intelligence Development Strategy Research Institute, Liu Xuan, researcher at the China Association for Science and Technology Innovation Strategy Research Institute and deputy director of the Innovation Environment Research Institute, as well as representatives from the Chinese Academy of Social Sciences, Peking University, Zhejiang University, Nankai University, Harbin Institute of Technology, Shanghai Academy of Social Sciences and More than 30 experts and scholars from Fudan University and other universities at home and abroad gathered in the Siyuan Hall of Guanghua Building of Fudan University to conduct in-depth discussions on the ethical challenges of artificial intelligence, especially generative artificial intelligence. From different perspectives of Eastern and Western philosophy, they focused on the theoretical basis, technology and ethical approaches of artificial intelligence philosophy and machine ethics.

The conference included six sub-forums and an ethics forum for outstanding international young scholars.

opening ceremony

The opening and closing ceremonies of this conference were hosted by Wang Guoyu, director of the China Association for Science and Technology-Fudan University Institute of Science and Technology Ethics and the Future of Humanity and professor of the School of Philosophy of Fudan University.

During the opening ceremony, Wang Guoyu introduced to the guests the background and topics of this conference, as well as the mission and research direction of the China Association for Science and Technology-Fudan University Institute of Ethics of Science and Technology and the Future of Humanity, as well as the achievements the institute has made in the past year in the "Four One Projects" - a magazine, a laboratory, an academic platform for internal and external exchanges, and the construction of an excellent research team. Wang Guoyu believes that the challenges brought by artificial intelligence involve both the contemporary era and the future of mankind. They are both local and cultural, and many of them transcend national boundaries and cultures. It is our shared responsibility to promote a truly inclusive global dialogue, guided by our shared values ​​and commitment to human dignity.

Speech session

Zhou Hu, deputy secretary of Fudan University, attended the meeting and delivered a speech. He pointed out that artificial intelligence technology is changing with each passing day, and its far-reaching impact has penetrated into all aspects of social life. It not only redefines the relationship between people and the world, and people, but also restructures the way we acquire, understand and apply knowledge on a deep level, and even reshapes human cognitive map to some extent. In March 2022, my country issued the "Opinions on Strengthening the Governance of Science and Technology Ethics", emphasizing that strengthening the theory and research of cutting-edge major issues in science and technology ethics and building a science and technology ethics governance system based on Chinese characteristics are important measures to ensure the healthy development of a community with a shared future for mankind. Fudan University has always been committed to the promotion of science and technology ethics in education, research and practice. The convening of this international academic conference on "Cross-Cultural Dialogue on Artificial Intelligence Ethics" is an important step for our school to further participate in global artificial intelligence ethics research topics. He hopes that this conference will build a continuously interactive international high-end academic exchange platform to jointly respond to global scientific and technological ethical challenges and help build a more responsible and sustainable scientific and technological future.

Zhang Shuangli, Dean of the School of Philosophy at Fudan University, delivered a speech and expressed a warm welcome to the guests. She pointed out that this event is an important part of Fudan Philosophy Subject Week, and the School of Philosophy has always been committed to supporting and developing high-quality academic platforms. Since its establishment more than a year ago, the China Association for Science and Technology-Fudan University Institute for the Future of Science and Technology Ethics has focused on key and difficult issues in science and technology ethics and is committed to promoting interdisciplinary and cross-cultural academic dialogue on science and technology ethics. It is expected that scholars at the conference will reflect deeply on ethical issues in science and technology and conduct high-level, cross-cultural academic dialogue.

first meeting

The theme of the first conference was “Artificial Intelligence Philosophy—Eastern and Western Perspectives”.

Professor Kaukolberg delivered a report titled "AI & Good: a Cross-and". He combined the different traditions of China and the West and connected the concept of common good (Good) with the development prospects of AI. He pointed out that establishing a bridge of communication between different cultures is of great significance to the ethics of contemporary AI technology.

Xu Yingjin, a professor at the School of Philosophy at Fudan University, published a report titled "and the of Data." He reinterpreted the economic implications of Confucian retro thought from the perspective of historical materialism. He pointed out that in the contemporary world, individual privacy rights are fully threatened by big data technology. The Confucian view of privacy based on clan consciousness, after experiencing a modern interpretation of the concept of "clan", may be able to effectively supplement the Western liberal view of privacy based on individuals. The combination of small data technology and home robot technology, as well as the reasonable use of blockchain technology, may provide a practical technical carrier for the revival of the above-mentioned Confucian view of information in the context of an aging population.

Bai Xue, associate professor at the Department of Philosophy and Law at the Perm National Research Polytechnic University in Russia, delivered a report titled "Human-Robot (HRI) in: Value and ' '". She pointed out that the introduction of robots into society has promoted the establishment of human-robot interaction (HRI) as an interdisciplinary research field. The ethical issues of human-computer interaction not only involve how to correctly apply a new technology, but also include the conflict and integration between different values. The rapid growth of robotics technology may bring potential security and privacy issues. It is expected that new solutions can be contributed through friendly cross-cultural exchanges and communication.

second meeting

The theme of the second meeting was "Technology and Context".

Professor Kernan delivered a report titled "as the of and an of". On the basis of analyzing the relationship between artificial intelligence technology and modernity, he pointed out that modernity is a trauma to mankind. Unfortunately, it still dominates our cooperation and unity in this world. Despite his pessimistic stance on the future of mankind, human cooperation and exchanges can still be seen amid disputes and disasters, and we can see everyone uniting to face challenges.

Wang Tianen, professor of the Department of Philosophy of Shanghai University, delivered a report titled "AI in". He pointed out that with the development of artificial intelligence, its ethical issues are becoming more and more important and special. On the one hand, the understanding of AI ethical issues is based on the understanding of AI itself and its development; on the other hand, the development of AI highlights the special situation of the development of ethics. Since the understanding of AI and its development increasingly must go deep into the information level, AI ethical issues can be better understood in the information context based on the understanding of the susceptibility relationship of information.

third meeting

The third session was titled “Ethical Considerations of Generative Artificial Intelligence”.

Professor Nordmann delivered a report entitled "Back and the End of". He pointed out that the author () means the original man, the creator of the world, the creator of texts and actions. The birth of "" allows us to see that the author's status is overthrown, but it occurs in the mundane and ordinary, in the redefined human-computer interaction, which makes a new set of values ​​​​and a new subject come to the fore. We ourselves are reconstructed when we see ourselves imitated in the mirror image of technology that replicates our behavior.

Duan Weiwen, a researcher at the Institute of Philosophy of the Chinese Academy of Social Sciences and a professor at the University of the Chinese Academy of Social Sciences, published a report titled "as a -Moral Guide and Its". He pointed out that generative AI is a new stage in the development of computational intelligence and digital thinking. Its essence is that a new generative space is automatically generated on top of the real space and data space. With the development of artificial intelligence navigation, "co-pilot" and other intelligent agents, digital data-driven agents are becoming a mechanism to guide human schedule behavior, making AI a guide in our lives, our knowledge and moral mentor, and bringing us a new lifestyle characterized by intelligent domestication.

Zhao Yanyan, professor of the Department of Computing at Harbin Institute of Technology, delivered a report titled "Large with Human". She pointed out that on the basis that the understanding ability of large language models is similar to human abilities, she asked: Can human value be calculated? Are human values ​​at the individual level or at the group level? What is the difference between values ​​and morals? Wait three questions. She introduced related tasks and algorithms for value alignment of large language models, that is, detection and alignment from three levels: laws and regulations, ethics, and psychological safety. Finally, she pointed out that large language models still face many challenges, such as conflicts between diverse human values ​​and the goals of "usefulness" and "harmlessness" of large language models.

Andres Kaminsky, professor of philosophy at the Technical University of Darmstadt in Germany, published an online report titled "Is It or Is It? with the of a Key in AI and Its for the of." Through analyzing four cases, he pointed out that modern technology does not necessarily develop completely in accordance with the designer's original design ideas, and it is difficult for people to imagine in advance the possible real consequences of each specific technology. The main reason for the gap between technical design and real-life consequences is the opacity of technology, that is, people's cognitive limitations make it difficult to imagine all possible future situations based on the current situation; the opacity of the relationship between the data, information and expected goals input during design leads to the above problems.

fourth meeting

The theme of the fourth conference is "Artificial Intelligence, Technology and Ethics".

Professor Grunwald published a report titled Al-as: The View on, in which he pointed out that the “disruption” brought about by artificial intelligence is gradually accumulated and therefore easily ignored. But at the same time, the destructive power of this "subversion" is huge. It will threaten human freedom and independence and damage the openness of the future. Therefore, we must be keenly aware of the potential ethical issues of artificial intelligence from now on and actively think about response strategies.

Researcher Cheng Sumei from the Institute of Philosophy of the Shanghai Academy of Social Sciences published a report titled The of. She pointed out that persuasive technology has been applied in many fields. Computing products containing persuasive functions may conceal their persuasive intentions, limit users’ freedom of choice, and induce users’ Internet addiction. To deal with these ethical challenges, we need to transcend the dichotomy between theoretical ethics and applied ethics, expand the semantic and pragmatic scope of the concept of responsibility, shift from passive and responsible technical ethics to active and responsible technical ethics, and establish an ethics about the future.

fifth meeting

The theme of the fifth conference is "Theoretical Basis of Artificial Intelligence Ethics".

Professor Mitchum published a report entitled Arranging Two AI. He pointed out that when we talk about the ethics of artificial intelligence, “ethics” alone is not enough. Starting from the ethical tradition starting from Aristotle, we can see that ethical analysis and political thinking are equally important for the ethics of artificial intelligence. At the same time, he believes that artificial intelligence ethics is only about "artificial intelligence" and it is not appropriate. If we want to properly deal with the negative impacts of artificial intelligence, we must rely on changes at the sociopolitical level.

Professor Liao Bishui from the School of Philosophy of Zhejiang University published a report titled, Basic, and of. He pointed out that the key way to solve the problem of how to give sufficient ethical considerations to the decision-making of intelligent agents is to establish machine ethics based on human ethics. Machine ethics has four basic characteristics that are different from human ethics: (1) Current intelligent machines show weak subjectivity in ethical decision-making; (2) Machine ethical decision-making needs to strike a balance between the value demands of different stakeholders, that is, it has social balance; (3) Machines are easily affected by cultural factors when making ethical decisions and should be able to reflect cultural differences; (4) Machines need to explain ethical decisions to human subjects, understand emotional expressions, and attribute responsibility, so they must have good human-computer interaction.

The picture shows Wang Haosheng, postdoctoral fellow of Professor Li Hengwei, giving a report on his behalf

Professor Li Hengwei from the School of Philosophy of Zhejiang University and others published a report titled Human-: The , , and . They pointed out that both strategies of effective accelerationism and decelerationism try to solve crises such as environmental damage and resource depletion through human control of technology. However, these two perspectives are too simplified. The metaphors of Prometheus and Epimetheus reveal the unpredictability in the development and use of technology, while the concept of technosphere urges us to rethink how to find balance in this complex system composed of humans and technology to achieve symbiosis between humans and technology. As a hybrid of human and machine, cyborg provides the possibility of realizing symbiosis between man and technology.

The sixth meeting

The theme of the sixth conference was "Ethics of Large Language Models".

Professor Xiao Yanghua from the School of Computer Science at Fudan University published a report titled "The of AIGC ( ) in the Era of: Basic and ". He pointed out that the rapid development and widespread application of artificial intelligence technology, especially AGI, will have a double impact on human society. On the one hand, AGI is expected to become an advanced productive force that promotes social progress. But on the other hand, the abuse of AGI may also bring a series of fundamental challenges to mankind. AGI's full penetration and replacement of human mental work will damage human core intellectual abilities and have an impact on the fundamental question of "what it means to be human." Therefore, AI technology should avoid harm to human nature and prevent AI abuse. Guidelines should be established for AI applications and should not be used excessively or without restrictions.

Professor Yan Hongxiu from the Shanghai Jiao Tong University Institute of Science History and Science Culture published a report titled "Value and Trust". She pointed out that in the process of aligning human values ​​​​with artificial intelligence values, deceptive value alignment completes the superficial value alignment with an amoral strategic trust. This kind of deception involves human-machine deception, human-human deception and machine-machine deception. With the emergence of learning from weak to strong and self-alignment, the study of deceptive value alignment is not only a problem that must be solved to build a good human-technology relationship, but also a necessary condition to ensure that artificial intelligence is good.

Professor Wang Guoyu delivered a report titled From Value to Value: Its and. She discussed the theoretical feasibility, implementation methods and limitations of "value alignment". On the basis of clarifying the connotation and importance of the concept of "value alignment", the report shared a number of original experimental studies conducted by Wang Guoyu's team. The conclusion of the study not only proves that the value alignment method of the current behaviorist approach can effectively improve the moral ability of the large language model, but also reveals the limitations of this approach.

International Ethics Forum for Outstanding Young Scholars

During the conference, an ethics forum for outstanding international young scholars was also specially arranged. Zhu Linfan, a young associate researcher at the China Association for Science and Technology-Fudan University Institute of Science and Technology Ethics and the Future of Humanity, presided over the youth forum.

Jia Xiangtong, a professor at the School of Philosophy of Nankai University, gave a report on "Not Be a Fifth in". He pointed out that it is necessary to add an interpretability principle to the principles of artificial intelligence ethics, and this principle should be integrated into the principle of openness to further consolidate the comprehensive framework of artificial intelligence ethics.

Qu Jingjing, associate researcher at Shanghai Artificial Intelligence Laboratory Governance Research Center, gave a report "for GPT with -". She pointed out that the advantage of using machine evaluation is its economy. According to its interview scale, she believed that China's large language model can effectively enhance the doctor-patient relationship. She also pointed out that real doctors make judgments on patients based on years of experience and command, and machines should be consistent with human command.

Teng Yan, a young researcher at the Shanghai Artificial Intelligence Laboratory Governance Research Center, gave a report on "Human into LLMs: An of, , and". She introduced human values ​​​​into value design and pointed out the ethical and technical challenges faced by this approach. Her team is evaluating and improving this.

Alex Andra Kazakova, a doctoral student at the University of Chinese Academy of Sciences, gave a report on "In AI We Trust? The on AI in". She described the picture of the top-down artificial intelligence revolution and pointed out that people will produce a fragmented self-projection when interacting with artificial intelligence.

closing ceremony

In his speech at the closing ceremony, Professor Wang Guoyu and domestic and foreign experts present imagined new developments and new challenges in the field of artificial intelligence technology and ethics in the future, and expressed his heartfelt thanks to the guests and support staff who attended the meeting. The convening of this international academic conference has effectively enhanced Fudan University's international influence in the field of artificial intelligence ethics and governance.