When generative AI enters the high-speed development track, its ethical governance issues have also attracted much attention from all walks of life. Based on analyzing the ethical problems brought about by generative AI, this paper systematically explains my country's ethical governance system.

Author丨Fan Xiaojuan Yin Lei Huang Haiyi

introduction

According to Beijing Youth Daily, during the Qingming Festival in 2023, a young man surnamed Wu from Shanghai used AI tools to generate a digital virtual person with his deceased grandmother and had a conversation with her. In the video, "grandma" speaks the dialect of Hubei, with gray hair, and "nagging" like she was before her lifetime. Just looking at the expression can be said to be vivid. [1] Some netizens expressed their feelings about this, thinking that this is a very meaningful way to express grief; while some netizens believe that making deceased relatives into AI is like making relatives into manipulated electronic dolls.

In recent years, generative AI products and services such as AI writing, AI painting, AI music, and AI game development have become popular. While the rapid development of generative AI empowers the real economy, related ethical issues are emerging one after another, and ethical governance has become an urgent topic.

1. Ethical issues

At present, the ethical issues caused by AI applications have attracted widespread attention worldwide. At the beginning of 2022, UNESCO published the "Recommendation on the Ethical Issues of Artificial Intelligence". The proposal mentions: “AI systems raise new ethical issues, including but not limited to their impact on decision-making, employment and labor, social, health care, education, media, access to information, digital divide, personal data and consumer protection, the environment, democracy, the rule of law, security and policing, dual use, human rights and fundamental freedoms (including freedom of expression, privacy and non-discrimination). In addition, AI algorithms may replicate and deepen existing biases, thereby aggravating existing forms of discrimination, bias and stereotypes, resulting in new ethical challenges.”

As an important part of AI technology, the ethical issues involved in generative AI include but are not limited to the following aspects:

(I) The lack of social interaction

Based on technologies such as deep synthesis, generative AI applications such as virtual people and chat robots provide users with interesting and simulated social methods, and also provide users with new and immersive companionship methods. Deep synthesis technology can already "resurrect" deceased relatives and express the longing of the living; Xiao Ai classmate, Tmall elves, etc. can also sing, tell stories to children, and answer their children's "100,000 whys"; "AI partners" can even "fall in love" with users. Users can communicate with other users in the metaverse, or simply chat with virtual people, and meet their emotional needs with virtual people.

Artificial intelligence can be used to accompany the elderly or children, and can reduce the burden on caregivers. But can artificial intelligence or machines replace family affection and family companionship? The answer is obviously no. According to Maslow's theory of hierarchy of needs, the need for love and belonging is second only to human physiological and safety needs. Leaving the elderly or children mainly to robot care may reduce their connection with society, and their emotional needs are difficult to meet, and may even increase the risk of illness.

(II) Discrimination and false information

The operation system of generative AI is based on deep learning. The basis of deep learning is that the R&D team artificially "feed" massive training data to AI and design algorithms for AI. The content generated by AI depends on the statistical rules in its training data, but due to the limitations of current technology, events that contain discriminatory information or false information in the content generated by AI still occur from time to time.

Taking discrimination as an example, the risk test results of DALL-E-2 generated content in April 2022 show that the image generation results may reflect more stereotypes and biases. For example, when entering words such as "week" and "restaurant", the program will be more inclined to generate pictures of Western scenes; when entering keywords such as "CEO" and "lawyer", DALL-E-2 also generates human images of male and white people more frequently [2]. Taking false information as an example, according to CNN, in the case of Mata v. Airline, the six cases cited by senior lawyers in the filing of court documents were found to be forged. The 6 cases used were from, but I did not verify the correctness and error of the information. [3]

(III) Labor substitution

Industrial automation replaces human muscles, while digital automation replaces human thinking or information processing. [4] Applications such as GPT-4, AI painting, and AI game development show that the use of generative AI will undoubtedly greatly improve productivity and productivity. However, increasing productivity and productivity through the use of generative AI also means that the same output requires only less manpower.

In the previous few industrial revolutions, the labor market has undergone tremendous changes. The advent of the digital age and the application of artificial intelligence and robots will cause more thorough changes in the labor market. The main question is: Will the consequences of this time be different? Can the new jobs generated and the new wealth created by them cover the adverse effects of the objective reduction of jobs? Even if there is no obvious change, what is the transition cost of this stage, and who will pay for it? Do we need to make mandatory social adjustments to distribute the benefits of digital automation fairly? [5]

2. Principles and positions of ethical governance

The Chinese government’s position and principles on the ethical governance of artificial intelligence are very clear and firm. On November 16, 2022, China submitted the "China's Position Document on Strengthening the Ethical Governance of Artificial Intelligence" (hereinafter referred to as the "Position Document") to the Conference of the Parties to the United Nations Convention on Specific Conventional Weapons. The Position Document pointed out that "China has always been committed to building a community with a shared future for mankind in the field of artificial intelligence, actively advocated the concepts of 'people-oriented' and 'intelligent goodness', and advocated enhancing the understanding of the ethical issues of artificial intelligence in various countries, ensuring that artificial intelligence is safe, reliable and controllable, better empowering global sustainable development, and enhancing the common well-being of all mankind." In terms of supervision, the Position Document put forward a number of important propositions, including: adhering to ethics first; suggesting the establishment and improvement of the ethical norms, norms and accountability mechanism of artificial intelligence; suggesting the establishment and improvement of the normative and legal policy system of artificial intelligence ethics; and enhancing bottom-line thinking and risk awareness.

On March 20, 2022, the "Opinions of the General Office of the CPC Central Committee and the General Office of the State Council on Strengthening the Governance of Science and Technology Ethics" clearly stipulates the principles of science and technology ethics as: "Improve human well-being, respect the right to life, adhere to fairness and justice, reasonably control risks, and maintain openness and transparency."

III. Legislation of ethical governance

In 2016, the "Measures for the Ethical Review of Biomedical Research Involving Humans" (hereinafter referred to as the "Measures") was released. The "Measures" carried out ethical review of biomedical research involving humans for medical and health institutions at all levels and types.

In order to adapt to the technological development of artificial intelligence, on the basis of the "Measures", on February 18, 2023, the National Health Commission and others issued the "Measures for the Ethical Review of Life Sciences and Medical Research Involving Humans" (hereinafter referred to as the "Measures for the Review of Measures"). The Review Measures extend the scope of application from the medical ethical review of the Measures to research activities carried out by human-as-subjects or using biological samples, information data (including health records, behaviors, etc.) of people (collectively, research participants), and conduct ethical review of them.

On April 4, 2023, the Ministry of Science and Technology issued the "Measures for Ethics of Science and Technology (Trial) (Draft for Comments)" (hereinafter referred to as the "Trial Measures"), aiming to further improve the ethical review of scientific and technological activities and make up for the gap in the ethical review provisions for scientific and technological activities outside the scope of medical ethics review.

The following is a comparison of the scope of application and applicable objects of the three documents.

Click to view the larger image

Activities within the scope of application of the corresponding provisions shall pass ethical review. If there is competition and cooperation in the scope of application, they shall pass different ethical reviews at the same time. If the corresponding activities that fail the ethical review may face administrative penalties and criminal liability.

4. Scientific ethics review

At this stage, China's main method of ethical governance is ethical review. With the release of the "Trial Measures", it has expanded the scope of ethical review from scientific and technological activities that directly target human research to all scientific and technological activities with ethical risks, filling the gap in the relevant provisions on ethical review of scientific and technological activities outside the scope of medical ethical review.

(I) Responsible subject for scientific ethics review

According to Article 5 of the "Trial Measures", colleges and universities, scientific research institutions, medical and health institutions, enterprises, etc. are the responsible entities for the review and management of science and technology ethics. Units engaged in scientific and technological activities such as life sciences, medicine, artificial intelligence, etc., whose research content involves sensitive areas of scientific and technological ethics, a science and technology ethics (review) committee should be established. Other units that have ethical review needs can establish a science and technology ethics (review) committee based on actual conditions.

(II) Scientific ethics review process

Scientific and technological activities within the scope of application must pass the scientific and technological review before they can be carried out. The general review shall be subject to the person in charge of science and technology activities to the unit’s Science and Technology Ethics (Review) Committee. The Science and Technology Ethics (Review) Committee shall decide whether to accept the acceptance and provide a written notice. In principle, the review method is meeting review. The Science and Technology Ethics (Review) Committee may make decisions such as approval, modification approval, retrial after modification or non-approval for scientific and technological activities under review. The review period is generally 30 days.

The "Trial Measures" also establish a list system for ethical high-risk scientific and technological activities. In addition to general review procedures, the technology on the list adds expert review procedures for local or industry authorities. The unit shall actively register through the National Science and Technology Ethics Management Information Registration Platform within 30 days after the scientific and technological activities included in the list are approved for ethical review and approval.

(III) Review focus

According to Article 14 of the "Trial Measures", the standards and focus of scientific and technological ethics review include the qualifications of scientific and technological personnel, research foundation and facility conditions; risk-benefit ratio of scientific and technological activities; rights protection of participants in scientific and technological activities; experimental animal welfare; data processing plans and data security risk control, etc.

(IV) Protection of the rights of research participants

Consistent with the Review Measures, the Trial Measures also focused on the protection of the rights of research participants. According to Article 14 of the "Trial Measures", the rights of research participants include fair and reasonable right to participate, personal information and privacy protection, right to compensation and damages, right to know, and special protection for vulnerable groups.

Conclusion

The premise for technological innovation and progress, ethics and compliance must adhere to the concept of first-hand. People should not only care about immediate interests, but the development of science and technology must always adhere to a people-oriented center. Only by adhering to the bottom line of ethics can generative AI better empower the real economy and reduce costs and increase efficiency for digital economic activities.

[Note]

[1] Beijing Youth Daily, Liu Meimei, "The young man used AI technology to "resurrect" his grandmother, netizens started arguing",

[2], last visit date: May 31, 2023

[3], last visit date June 1, 2023

[4] C. Müller, of and, published in of, last visited: April 12, 2023.

[5] C. Müller, of and, published in of, last visited: April 12, 2023.

statement

This article is intended to be a general analytical research or information sharing of regulations, and does not constitute any result of analytical research and judgment of specific laws, nor does it serve as any basis for any advice or advice provided to readers. The author hereby expressly disclaims liability for any action or omission taken in accordance with this article. If you need to reprint or quote any content in this article, please contact the author (); No content in this article may be reproduced or quoted without the author's consent.

Fan Xiaojuan Lawyer

Shanghai Office Partner

Business areas: Private equity and investment funds, Hong Kong and overseas capital markets, financial products and trusts

Featured Industry Category: Communications and Technology

Yin Lei Lawyer

Shanghai Office Private Equity Fund and Asset Management Department

Huang Haiyi Lawyer

Shanghai Office Private Equity Fund and Asset Management Department

Special statement