The development of artificial intelligence has always gone hand in hand with ethics. The rapid iteration of artificial intelligence technology has promoted the improvement of my country's science and technology ethics system. Recently, the Ministry of Industry and Information Technology, together with multiple departments, drafted the "Artificial Intelligence Technology Ethics Management Service Measures (Trial) (Draft for Public Comments)" (hereinafter referred to as the "Measures"), whereby scientific and technological ethics norms are deeply embedded in artificial intelligence at the organizational and technical levels. The "Measures" embody Chinese wisdom in the construction of an artificial intelligence regulatory system and propose a model plan for the ethical governance of artificial intelligence. Combining industry and technology trends, the expansion of ethical norms and ethical subjects to individual users and artificial intelligence itself can be explored in the future.

Science and technology ethics are embedded in the organizational structure of carrying out artificial intelligence science and technology activities. Article 9 of the "Measures" stipulates: "Universities, scientific research institutions, medical and health institutions, enterprises, etc. that are engaged in artificial intelligence science and technology activities are the responsible entities for the ethics management services of artificial intelligence science and technology in their units. Units with conditions shall establish an artificial intelligence science and technology ethics committee." The artificial intelligence science and technology ethics committee (hereinafter referred to as the committee) includes experts with backgrounds related to artificial intelligence technology, applications, ethics, law, etc. Its main responsibilities include guiding scientific and technological personnel to carry out scientific and technological ethics risk assessments, tracking and supervising the entire process of relevant scientific and technological activities as required, organizing and carrying out scientific and technological ethics knowledge training, etc. Following this, organizations that carry out relevant scientific and technological activities can internalize scientific and technological ethics in their organizational structure and activity processes by establishing committees and exerting their guidance, supervision, and training responsibilities.

Technological ethics are embedded in every aspect of the life cycle of artificial intelligence systems. Article 15 of the "Measures" stipulates: The ethical review of artificial intelligence focuses on five aspects: fairness and justice, controllability and credibility, transparency and explainability, traceability of responsibilities and personnel qualifications. Among them, the first four aspects cover the integrity and diversity of the data, the rationality of the algorithm model and system design, the robustness of the system design and the continuous monitoring of the system operation, etc., and require "taking measures such as log management to clearly record the data, algorithms, models, and relevant information of each link of the system to ensure that the entire link can be tracked and managed, and when problems occur, specific links can be accurately located and the responsible person can be determined." In other words, from collecting data during development and designing the system to putting it into application and continuing operation after completion, all aspects of the artificial intelligence system must comply with scientific and technological ethics.

Science and technology ethics are deeply embedded in artificial intelligence at the organizational and technical levels, reflecting China’s stance on the construction of an artificial intelligence normative system. From a global perspective, there has been no generally accepted answer for a long time on how to place technological ethics in the artificial intelligence normative system and how to deal with the relationship between technological ethics and legal norms. In 2024, the EU Artificial Intelligence Law will be officially passed. Article 25 of its preamble clearly states: On the one hand, the ethical principles of artificial intelligence help to design coherent, trustworthy, and human-centered artificial intelligence systems; on the other hand, they cannot hinder the implementation of the Artificial Intelligence Law and other EU laws. In other words, law takes precedence over ethics, and no ethics review or similar system should be set up in special legislation on artificial intelligence. The "Measures" reflect another way of thinking, paying more attention to the synergy and two-way connection between scientific and technological ethics and legal norms. In the future, as scientific and technological ethics become more important in the era of artificial intelligence, the "Measures" may provide a model for the global governance of artificial intelligence.

Combining industry and technology trends, we can explore the expansion of scientific and technological ethics norms to individual users. At present, institutions of higher learning, scientific research institutions, medical and health institutions, enterprises and other units are the entities responsible for the ethics of science and technology. As artificial intelligence becomes increasingly integrated into social life, the scenarios and frequency in which individual users use artificial intelligence increase. The likelihood and severity of harm to itself and other users also increases accordingly. If an individual user suffers serious damage due to addiction to artificial intelligence, the intervention of ethical standards is likely to be required; and if an individual user uses artificial intelligence to "resurrection" to commemorate the deceased, it may lead to more complex ethical issues, etc. In the era of artificial intelligence, it is even more necessary for individuals to have certain scientific and technological ethics. We can start with literacy cultivation and gradually establish rules in a flexible way in scenarios where there are already recognized ethical judgments.

Combining industry and technology trends, we can explore the expansion of scientific and technological ethics subjects into artificial intelligence itself. Currently, artificial intelligence systems are only one of the subjects of ethical review. For example, Article 21 and Annex 1 of the "Measures" stipulate that for automated decision-making systems with a high degree of autonomy that are oriented to scenarios involving safety, personal health risks, etc., on the basis of a preliminary review of science and technology ethics, relevant experts also need to conduct a review to ensure that they meet the requirements in all aspects such as fairness, justice, controllability and credibility. Since 2024, research on the welfare of artificial intelligence has begun to appear internationally. In other words, recognize the ethical subject status of artificial intelligence and explore ways to define and respect its well-being. A few cutting-edge artificial intelligence companies have begun to implement welfare assessments before releasing artificial intelligence, and even allow artificial intelligence to actively terminate conversations with users. Of course, research on the welfare of artificial intelligence has just begun. On the other hand, if conclusions are reached, it may have a great impact on policy, industry, and academic science. In short, it is worthy of in-depth study. It is not mature yet and needs more discussion.

Author: Zhu Yue, assistant professor at Tongji University Law School and researcher at Shanghai Artificial Intelligence Social Governance Collaborative Innovation Center