The Paper reporter Zhou Dishaowen

In recent years, the artificial intelligence (AI) industry has developed rapidly, and its ethics and governance issues have also attracted widespread attention. How to understand artificial intelligence ethics in different cultural backgrounds? What kind of cognitive logic and framework paradigm for artificial intelligence governance should be constructed? How can artificial intelligence ethics move from theory and concept to practice?

"New technologies often have an impact on existing social ethical norms, and artificial intelligence technology is no exception. This is a problem that human society has always faced in the process of progress." On June 1, Xue Lan, director of the National New Generation Artificial Intelligence Governance Professional Committee, senior professor of liberal arts at Tsinghua University, and director of the Artificial Intelligence International Governance Institute, said at the Artificial Intelligence Ethics, Governance and Sustainable Development Forum of the 2022 Beijing Intelligent Source Conference.

Xue Lan proposed that we should build a communication platform to solve information asymmetry, seek common ground while reserving differences, and find a solution mechanism that is consistent with the development process of the entire society. At the same time, we should also set up an update mechanism for technological and conceptual updates.

In recent years, my country has successively issued policy documents such as the "New Generation Artificial Intelligence Governance Principles" and the "New Generation Artificial Intelligence Ethical Code", clearly proposing eight principles and emphasizing the integration of ethics and morality into the entire life cycle of artificial intelligence. Zhao Zhiyun, Secretary of the Party Committee and Director of the China Institute of Scientific and Technological Information of the Ministry of Science and Technology, and Director of the New Generation Artificial Intelligence Development Research Center of the Ministry of Science and Technology, said that the Artificial Intelligence Subcommittee of the National Science and Technology Ethics Committee is studying and drafting a high-risk list of artificial intelligence ethics to better promote the ethical supervision of artificial intelligence scientific research activities.

Safety and legal issues are the bottom line and should not be generalized into the scope of ethical governance

According to incomplete statistics, more than 100 documents related to the ethical governance of artificial intelligence have been released around the world. Governments, international organizations, academia and industry are paying close attention to the issue of ethical governance of artificial intelligence.

"Governance and development should not be opposed, but governance should be used to better protect the development of artificial intelligence and strive to achieve a balance between high-quality development of artificial intelligence and high-level governance." Zhao Zhiyun said that ethical governance provides benign guidance for artificial intelligence activities and is an important guarantee for the healthy development of artificial intelligence.

Zhao Zhiyun said that there are three key requirements for promoting ethical governance of artificial intelligence, namely, grasping the development laws and characteristics of artificial intelligence and clarifying practical ethical issues; grasping the boundaries of ethical governance of artificial intelligence and regulating the development of artificial intelligence within an appropriate scope; grasping the balanced relationship between development and governance, and promoting development through governance.

She pointed out that current artificial intelligence technology has not yet reached the level of strong artificial intelligence and general artificial intelligence. Therefore, future ethical risks cannot be regarded as real risks to avoid unnecessary negative impacts on the development of artificial intelligence. In addition, it is necessary to clarify the bottom line requirements of artificial intelligence governance. Ethical governance mainly emphasizes the higher demand of promoting innovation and preventing risks. Safety and legal issues are the bottom line that cannot be crossed and should not be generalized into the scope of ethical governance.

Build an artificial intelligence evaluation platform (system) to help implement artificial intelligence governance

Talking about the social attributes of the development and application of artificial intelligence technology, Li Renhan, a member of the National New Generation Artificial Intelligence Governance Professional Committee and chief consultant professor of the Shanghai Jiao Tong University Artificial Intelligence Research Institute, attributed it to three aspects: safety issues, ethical issues, and mutation safety issues.

Security issues mainly include national security, social security, economic security, personal security, etc. Ethical issues mainly include privacy, fairness, transparency, discrimination, employment, etc. These two types of security issues have already occurred. As for mutation safety, that is, the problem that artificial intelligence may be uncontrollable and cause harm to society and humans, it has not happened yet and is an unconventional issue.

Li Renhan said that artificial intelligence is a technology that uses machines to simulate human cognitive abilities, and the requirements for it need to be the same as the requirements for humans in order to live in harmony. For regulating the governance of artificial intelligence. We should start from the aspects of standards, rules, laws, supervision, law enforcement, etc.

At present, the formulation of artificial intelligence standards by relevant international organizations has entered a blowout period, and the international artificial intelligence governance work certificate is accelerating to become operational. In this regard, Li Renhan believes that there is an urgent need to establish a corresponding rule system, which on the one hand will promote the sustainable and healthy development of AI, and on the other hand will help integrate with international standards and achieve global development. At the same time, an artificial intelligence evaluation platform (system) must also be built to help implement artificial intelligence governance.

Li Renhan said that the artificial intelligence evaluation platform (system) will become one of the important infrastructures for future social development and the basic guarantee for the artificial intelligence industry to empower the sustainable and healthy development of all walks of life. The platform (system) will also be an important stage for the cultivation, growth and play of relevant talents and an important basis for participating in the formulation of international rules. Therefore, it is necessary to plan as early as possible and implement the evaluation platform (system) as soon as possible in a "three-step" strategy.

Integrating ethics into the entire life cycle of artificial intelligence

In recent years, ethical controversies brought about by the application of AI technology have become more frequent: American research institutions published a paper stating that AI has designed biological robots that can reproduce themselves; previously in 2019, many companies such as Amazon used AI monitoring systems to monitor

Employees are evaluated and dismissed; and the increasingly popular application of facial recognition technology has also triggered widespread social concern about personal information security and other issues.

Regarding the issue of artificial intelligence governance, Xue Lan believes that the direct and indirect benefits and costs brought by the application of artificial intelligence technology to economic and social operations should be considered, and the pros and cons should be weighed. It is necessary to consider the impact of the application of artificial intelligence technology on various codes of conduct in society. In addition, the long-term economic and social impact of the application of artificial intelligence technology must also be considered.

Xue Lan suggested that we should improve the governance system of artificial intelligence ethics and safety; promote the implementation of laws, regulations and standards; establish a dialogue mechanism with multi-subject participation, track the frontiers of technological development, and discover new safety and ethical issues; establish ethical assessment, early warning and control mechanisms for cutting-edge research and application of artificial intelligence; strengthen ethical education for all relevant interest groups in artificial intelligence research and application, and improve ethical awareness; actively cultivate healthy intelligent culture and digital humanities ideas; strengthen scientific dissemination and ethical popularization of cutting-edge advances in artificial intelligence, and improve public awareness.

"Ethics must be integrated into the entire life cycle of artificial intelligence, provide ethical guidance for natural persons, legal persons and other relevant institutions engaged in artificial intelligence-related activities, and promote the development of artificial intelligence." Xue Lan said.

Talking about integrating ethics into the full life cycle of artificial intelligence, Zhao Zhiyun said that the government’s responsibility is to guide all types of entities to pay attention to ethical norms and embed them in operations to fully embody the concept of the full life cycle. For example, she said that in 2019, the Ministry of Science and Technology approved 18 cities as pilot zones for the new generation of artificial intelligence innovation and development. An indispensable part of the construction of the pilot zones is to conduct tests on ethical governance standards. In each pilot zone plan, there are corresponding operating standards for how various entities implement ethical norms.

Artificial Intelligence Transformation and Innovation Model and Social Governance Model

What are the essential differences between ethical issues in the field of artificial intelligence and ethical issues in other fields? What requirements does this difference place on the ethical governance paradigm of artificial intelligence?

In this regard, Chen Xiaoping, director of the Robotics Laboratory of the University of Science and Technology of China, director of the CAAI Artificial Intelligence Ethics Governance Working Committee, and executive member of the Global Artificial Intelligence Council, said that the most fundamental difference is the technology itself. In the past, automation technology could only replace part of human labor, but now artificial intelligence, even just specialized artificial intelligence, has far surpassed and replaced humans in specialized fields. Some philosophers predict that, except for those who work on artificial intelligence and those who make decisions, 97% of the jobs of other industry practitioners will be replaced. The fundamental reason lies in technology.

When talking about the differences between ethics and other industries, Chen Xiaoping made three points.

The first is the issue of controllability. If general artificial intelligence and strong artificial intelligence are realized in the future, perhaps all industry fields can be replaced by artificial intelligence. Chen Xiaoping said that although it seems unlikely at present, it cannot be ruled out theoretically. Once artificial intelligence has general capabilities, it may be able to achieve survival independent of humans, and this must be taken seriously.

The second is the challenge of artificial intelligence to innovation models. Previously, human technology was not powerful enough to challenge the innovation model, but now even specialized artificial intelligence has profoundly changed the innovation model.

The third is the issue of the integration of artificial intelligence and social governance in the new technological era. As artificial intelligence technology becomes more widely used, human society has entered an essentially new era. The traditional social governance model is no longer fully effective, and new social governance models must be developed.