As artificial intelligence increasingly penetrates into every aspect of our daily lives, and while human society is accelerating towards intelligence and digitization, technological ethics issues are also emerging one after another. From defeating the human Go world champion to privacy and security issues caused by facial recognition, to the difficulty of accountability due to frequent Tesla self-driving accidents, and workplace anxiety caused by the popularity of virtual humans, many scholars have expressed concerns about ethical governance issues.

Recently, the "Opinions on Strengthening the Governance of Science and Technology Ethics" (hereinafter referred to as the "Opinions") issued by the General Office of the Central Committee of the Communist Party of China and the General Office of the State Council pointed out that the focus should be on strengthening research on science and technology ethics legislation in the fields of life sciences, medicine, artificial intelligence and other fields. This is the first national-level guiding document for ethical governance of science and technology issued by our country, setting a "traffic light" for ethical governance of emerging technologies.

The real value of artificial intelligence is not to replace people

"Congratulations to 'Cui Xiaopan' for winning the Best Newcomer Award at Vanke Headquarters in 2021! As Vanke's first digital employee, 'Cui Xiaopan' officially 'joined' on February 1 this year... The write-off rate of overdue prepaid and receivable documents she urged reached 91.44%." At the beginning of this year, a WeChat message posted by Yu Liang, chairman of the board of directors of Vanke Group, flooded Moments. It is reported that this virtual person named "Cui Xiaopan" has been working for many days, and many Vanke employees did not even know that the colleague with whom he exchanged emails every day was not a real person. It was not until the "Best Newcomer Award" was announced that everything became clear.

With the rapid development of artificial intelligence technology, virtual humans are rapidly moving toward commercialization. And seeing virtual people with special skills entering the workplace, "workers" are increasingly worried about whether their "jobs" will be taken away. You must know that these virtual people can not only achieve "996", but also "007", and can even do it all year round without even paying a salary.

However, just like the two sides of the coin, the emergence of artificial intelligence has also filled the vacancies in some high-risk positions, such as jobs with toxic or hazardous working environments, or jobs that require long-term stay away from urban areas. Artificial intelligence has played a very good supplementary role in human resources. On the one hand, this has helped many companies solve the problem of "labor difficulties"; on the other hand, it has effectively alleviated the labor shortage under the increasingly serious aging trend.

As CEO and CTO of Space Exploration Technology Company () and CEO of Tesla, Musk’s attitude towards artificial intelligence is complicated. On the one hand, he warned that the "derailed" development of artificial intelligence is one of the three major threats currently facing mankind, and that it is necessary to avoid the thorough development of artificial intelligence; on the other hand, he is actively developing human-friendly artificial intelligence, seeking a symbiosis between humans and artificial intelligence through brain-computer interface technology, and almost all of his companies are related to artificial intelligence.

“The debate over the ethics of artificial intelligence explores the relationship between humans and machines, and more importantly, the relationship between humans using artificial intelligence as a medium.” Duan Weiwen, director of the Center for Science, Technology and Society Research of the Chinese Academy of Social Sciences, told a reporter from China Electronics News: “Currently, ethical governance of artificial intelligence mainly involves four major difficulties: First, controllability The second is reliability, accuracy and stability; the third is explainability and transparency; and the fourth is accountability. "In his view, every step in the development of machine intelligence is learning from human intelligence and cannot be separated from the help of human intelligence. In fact, human intelligence is smarter and smarter, not machine intelligence.

"The real value of artificial intelligence is not to replace people, but to put people first and benefit people. The relationship between artificial intelligence technology and ethics is not a relationship where one is stronger than the other, or you are strong and I am weak, but complementary to each other." Zhang Hui, director of the Megvii Artificial Intelligence Governance Research Institute, was interviewed by China Electronics News The reporter said in an interview, "In fact, for any technology, it cannot develop disorderly without the constraints of laws, regulations, social ethics, industry norms and other constraints. In this regard, artificial intelligence is no exception. Within a legal, compliant, and reasonable framework, artificial intelligence will only develop better and healthier."

Global artificial intelligence ethical governance is trending towards the same frequency

At present, at least more than 60 countries around the world have formulated and implemented artificial intelligence governance policies. It can be seen that the rules and order in the field of artificial intelligence worldwide are in the formation stage, and the development of ethical governance tends to be at the same frequency. Duan Weiwen pointed out: "The ethical governance of robots in foreign countries should be the most mature, and the United Kingdom, Singapore, and the European Union have relevant ethical design standards." For example, the United Kingdom has introduced the first design standard on robot ethics in history - "Guidelines for the Ethical Design and Application of Robots and Machine Systems." The Financial Stability Board (FBS) of the United Kingdom has formulated regulations for the application of artificial intelligence and machine learning in financial services, emphasizing reliability, accountability, transparency, fairness, and ethical standards.

The United States hopes to ensure and enhance its dominant position in the field of artificial intelligence, so it places more emphasis on the scientific nature and flexibility of supervision, and also attaches more importance to scientific and technological ethical governance in practical application fields, such as strict management of the use of large-scale biometric identification technology. The U.S. Securities and Exchange Commission (SEC) requires companies to delete some facial databases, including related algorithms.

The EU's regulatory style tends to be tough, and it has successively issued guiding documents such as the "EU Artificial Intelligence", "Trusted AI Ethical Guidelines", and "Algorithm Responsibility and Transparent Governance Framework", hoping to reshape the global digital development model through high-standard legislation and supervision.

Foreign technology companies such as Microsoft and Google are also actively exploring ethical governance of artificial intelligence. For example, Microsoft has three major organizations, including the Office of Responsible Artificial Intelligence (of AI), the Artificial Intelligence, Ethics and Engineering Research Committee (AI and and), and the Responsible AI Strategic Management Team (AI), which are respectively responsible for AI rule formulation, case studies, implementation supervision, etc., and have developed a series of technical solutions.

Google has stipulated the principles for the design and use of artificial intelligence from both positive and negative aspects, and promised to be willing to adjust these principles over time. It has also established a central team for responsible innovation to promote the implementation of ethical governance practices. For example, in order to avoid aggravating algorithm unfairness or bias, it has suspended the development of credit-related artificial intelligence products; based on technical issues and policy considerations, it has refused to pass facial recognition review proposals; research involving large language models continues with caution and cannot be officially launched before a comprehensive review of artificial intelligence principles.

In terms of artificial intelligence ethics research, foreign countries are more active. Stephen Schwarzman, co-founder, global chairman and CEO of the global private equity giant Blackstone Group in the United States, has donated US$188 million to Oxford University to fund research on the ethics of artificial intelligence. Musk also donated $10 million to the Future of Life Institute (of Life) to teach robots "ethics and morality."

Zhang Hui pointed out: "Governments, enterprises, relevant social organizations, and industry organizations in various countries are gradually strengthening dialogue and cooperation, and the United Nations is also playing an increasingly important role in the formulation of international artificial intelligence governance rules. It can be said that artificial intelligence governance has become a global consensus, and has entered the stage of establishing regulations and implementation from the conceptual level."

"Another point worthy of attention is that the dynamic governance of artificial intelligence technology will in turn promote the development of artificial intelligence technology itself." Duan Weiwen said. For example, the ban on face recognition technology in some scenarios is only temporary, and increased governance will promote related technical shortcomings to be filled faster.

Artificial intelligence ethical governance is still a new proposition

Artificial intelligence has had a profound impact on the entire human society. Zhang Hui said: "The greater the capability of technology, the more it needs to be properly managed. Artificial intelligence is not only an important means for us to create a better life, but also an object that needs to be properly managed."

Our country is at the forefront of the world in the field of artificial intelligence technology and is also a frontier explorer in the practice of artificial intelligence ethical governance. Some good practices are also being discussed and learned overseas. For example, at the policy level, the promulgation of the "Guiding Opinions on Strengthening the Comprehensive Governance of Internet Information Service Algorithms", "New Generation Artificial Intelligence Ethical Code", "Opinions on Strengthening the Ethical Governance of Science and Technology", etc. have provided top-level design guidance for the governance of science and technology ethics, and subdivided normative requirements have also appeared in some rapidly developing fields.

At the same time, some domestic technology companies such as Megvii, JD.com, and iFlytek have successively established specialized organizations such as AI Ethics Committees and AI Governance Research Institutes to vigorously promote the implementation of AI governance work within the company and itself. Tencent Research Institute and Tencent AI Lab jointly released the artificial intelligence ethics report "Technological Ethics in the Intelligent Era - Reshaping Trust in the Digital Society", advocating the establishment of new technology ethics for artificial intelligence. Ali Security explores "using AI to govern AI" to promote the sustainable development of artificial intelligence. Megvii was the first to publish the "Artificial Intelligence Application Guidelines" and has launched the selection of "Top Ten Global Artificial Intelligence Governance Events" for three consecutive years.

However, as of now, artificial intelligence ethical governance is still a new proposition, and there is not much historical experience to refer to. The entire industry is also "governing while developing", gradually improving laws and regulations and building a consensus on ethics.

"Artificial intelligence companies are facing a 'technical trap'." Duan Weiwen pointed out, "While artificial intelligence technology empowers society, it will also have an impact on social values ​​and ethics, and in the short term, the negative impact may be amplified. If some companies do not pay attention to ethical issues in their early development, it will lead to social distrust of them. For example, changing its name to Meta and transforming into the metaverse has attracted a lot of doubts, especially criticism of privacy and security. At this time, an 'ethical callback' is needed to make artificial intelligence 'trustworthy' through a series of corporate behaviors."

He believes that ethical issues brought about by the development of science and technology must be proactively responded to. For example, on data ethics issues, we need to strengthen data privacy protection and information security protection; on algorithm ethics issues, we need to urge platforms to eliminate bias and discrimination and maintain fairness and equity as much as possible; on information dissemination ethics issues, content platforms need to think about how to ensure content health and reduce false information when using artificial intelligence technology to distribute content. "It is not enough to formulate ethical norms. Artificial intelligence companies should take these measures to grasp a 'degree' and seek (social license), so that they can find the greatest common denominator between technology and ethics."

Finding the “greatest common denominator” must be a collaborative process involving multiple parties. Zhang Hui said: "The application chain of artificial intelligence is very long, including technology providers, system integrators, application software developers, individual developers and other enterprises and practitioners, as well as users and beneficiaries of AI products and applications in different industries. At present, various national laws and regulations are gradually improving the rights and responsibilities of different relevant parties. With the development of the industry, I believe that the ethical governance path of artificial intelligence will become clearer in the future. Enterprises must also strictly abide by laws and regulations in every process and link of promoting AI technology innovation and application. This is the red line and bottom line of all work."