The Ethical Path for Building a Responsible AI Governance System

The surging wave brought about by the rapid development of artificial intelligence technology, while profoundly changing human existence and lifestyles and bringing great convenience to humanity, has also broken the boundaries of traditional human-machine relationships, presenting new ethical challenges and moral dilemmas for human existence and interaction. How to better respond to and adapt to the uncertainties brought by the transformation of AI technology is an important issue we must face. General Secretary Xi Jinping pointed out: “Technology is a sharp tool for development, but it may also become a source of risk. We must prospectively assess the rule conflicts, social risks, and ethical challenges brought by technological development, and improve relevant laws and regulations, ethical review rules, and regulatory frameworks.” [1]1 Xi Jinping, “Speech at the 30th Collective Study Session of the Political Bureau of the 20th Central Committee of the CPC,” October 2023. This is a reference to a high-level Party meeting where Xi addressed the dual nature of technological progress. The “Governance Principles for a New Generation of Artificial Intelligence—Developing Responsible Artificial Intelligence” issued by the National New Generation Artificial Intelligence Governance Professional Committee explicitly proposes to “develop responsible artificial intelligence,” providing us with an important ethical path for the development and governance of AI. Further elucidating the ethical stance and ethical principles of responsible AI governance has become a necessary part of applied ethics research.

人工智能技术快速发展所掀起的澎湃汹涌浪潮，在深刻改变人的存在和生活方式，给人类带来极大便利的同时，也打破了传统人机关系的界限，给人类的存在和交往带来了新的伦理挑战和道德困境。如何更好回应和适应人工智能技术变革带来的不确定性，是我们需要面对的重要课题。习近平总书记指出：“科技是发展的利器，也可能成为风险的源头。要前瞻研判科技发展带来的规则冲突、社会风险、伦理挑战，完善相关法律法规、伦理审查规则及监管框架。”国家新一代人工智能治理专业委员会发布的《新一代人工智能治理原则——发展负责任的人工智能》明确提出要“发展负责任的人工智能”，这为我们发展和治理人工智能提供了重要的伦理路径。进一步阐发负责任人工智能治理的伦理立场和伦理原则成为应用伦理学研究的题中应有之义。

Ethical Challenges and Moral Dilemmas of Artificial Intelligence

人工智能的伦理挑战和道德困境

Artificial intelligence is triggering the most important technological revolution and social transformation, while also reshaping humanity itself and altering human behavior. With the rise of the biotechnology revolution and artificial intelligence, humanity is undergoing a dual technological transformation of body and mind, facing the risk of being replaced by technology, which raises new existential questions. In particular, with the new revolution in AI technology—the introduction of generative artificial intelligence (AIGC)—the boundaries of traditional human-machine relationships have been broken, manifested as “the anthropomorphization of AI” and “the mechanization and algorithmization of humans,” posing significant challenges to traditional ethical relationships and ethical systems.

人工智能正在引发最重要的技术革命和社会变革，同时也在改造着人类自身，改变着人的行为。随着生物技术革命和人工智能的兴起，人类正经历身体和精神的双重技术化，面临被技术替代的风险，引发新的存在论问题。特别是随着人工智能技术的新革命——生成式人工智能（AIGC）的推出，打破了传统人机关系的界限，体现为“人工智能的拟人化”和“人的机器化和算法化”，给传统的伦理关系和伦理体系带来巨大挑战。

With the advent of the era of artificial intelligence, the issues of “future humans” and “distant strangers” have gradually entered humanity’s ethical horizon. Traditional ethics emphasizes a “nearby ethics,” focusing on direct interaction between people, where both parties are familiar and co-present, and moral time and space are close-range. The deep involvement of technology has altered the moral space and nature of human behavior, introducing new objects, such as artificial intelligence, into the domain for which we are responsible and morally concerned. As a new technology, AI has brought profound changes to humanity’s ethical world and moral space, establishing new interactive relationships with humans in virtualized and anthropomorphized ways, expanding ethical relationships from “human-to-human,” “human-to-animal,” and “human-to-nature” to “human-to-AI.”

随着人工智能时代的到来，“未来人”和“远方的陌生人”问题逐渐进入了人类的伦理视界。传统伦理学强调的是一种“近的伦理学”，强调人与人之间的直接交往，行为者双方是熟悉的、共同在场的，道德时空是一种近距离的。而科技的深度介入，使人类的道德空间和行为性质发生了改变，一些新的对象，如人工智能，逐渐进入我们为之负责和道德关切的领域。人工智能作为一种新科技，给人类的伦理世界和道德空间带来了深刻变革，它以虚拟化和拟人化的方式与人类建立了新的交往关系，伦理关系从“人与人”“人与动物”“人与自然”扩展到“人与人工智能”。

Explaining the ethical responsibility relationship between humans and artificial intelligence using traditional human-to-human ethical relationships will face considerable difficulties, giving rise to a triple moral dilemma in human-machine ethical relations. The first is the good-evil dilemma. Traditional human-machine relationships are those between humans and objects, where technology serves as a tool used by humans. With continuous technological innovation, modern technology is no longer a simple tool, making it difficult for humans to distinguish whether the use of modern technology is good or evil. AI can both maximize benefits for humanity and, if misused, cause great harm. The second is the dignity dilemma. With the development of AI, especially the application of generative AI, excessive reliance on technology and AI may weaken human subjectivity and challenge human dignity. The third is the responsibility dilemma. Although AI currently does not possess genuine autonomy or subjectivity, it does exhibit certain autonomous behaviors. These autonomous behaviors bring ethical risks and uncertainties, leading to unclear boundaries of responsibility and issues of distributing ethical risks, sparking debates over the responsibility gap in AI applications and creating a vacuum in the traditional responsibility ethics system. How to resolve these moral dilemmas? Developing “responsible artificial intelligence” is undoubtedly a feasible ethical path.

人与人工智能之间的伦理责任关系，用传统的人与人之间的伦理关系来解释将面临较大的困难，会出现人机伦理关系的三重道德困境。第一重是善恶困境。传统的人机关系是一种人和物的关系，技术作为工具为人类所使用。随着技术的不断革新，现代技术不再是简单的工具，人类很难区分现代技术的使用是善的还是恶的。人工智能既可以最大限度地给人类带来福祉，也可能被滥用给人类带来巨大危害。第二重是尊严困境。随着人工智能的发展，特别是生成式人工智能的应用，人类过度依赖技术和人工智能，有可能会弱化人的主体性，挑战人之为人的尊严。第三重是责任困境。虽然目前人工智能还不具有真正意义上的自主性和主体性，但却具有了一定的自主性行为。这种自主性行为带来了伦理风险和不确定性，存在责任边界不明确和伦理风险的分配问题，引发了人工智能应用的责任鸿沟争论，使得传统的责任伦理体系出现了真空地带。如何解决这些道德困境，发展“负责任的人工智能”无疑是一种可行的伦理路径。

Adhering to a People-Centered Ethical Stance

坚持以人为本的伦理立场

Although the rise of artificial intelligence has opened up the horizon of ethical concern, extended traditional ethical relationships, and brought certain challenges and risks, it is certain that humans themselves are the deciders and users of AI development, and fundamentally, humans are the true responsible agents. “Responsibility” must run through the research, development, and application of AI. The “Governance Principles for a New Generation of Artificial Intelligence—Developing Responsible Artificial Intelligence” states that it should follow the principles of “harmony and friendliness, fairness and justice, inclusiveness and sharing, respect for privacy, safety and controllability, shared responsibility, open collaboration, and agile governance,” which undoubtedly provides us with a governance framework and action guide for resolving the triple dilemma. “Developing responsible artificial intelligence” means that humans research, develop, and use AI in a safe, reliable, and ethical manner. The connotation of “responsibility” includes inclusiveness, fairness, people-centeredness, transparency and explainability, safety, respect for privacy, and accountability, to ensure the responsible operation and use of AI systems. “Responsible artificial intelligence” does not mean developing artificial moral machines with moral and responsibility awareness that can autonomously think and make moral decisions and be responsible for their own actions. On the contrary, “responsible artificial intelligence” primarily points to humans being responsible; humans must shoulder the responsibility for the AI they create and develop AI responsibly so that it better serves humanity. Humans themselves are the key to AI development and governance, so we must adhere to a people-centered ethical stance.

尽管人工智能的兴起打开了伦理关切的视界，延伸了传统的伦理关系，带来了一定的挑战和风险，但可以确定的是人类自身是人工智能发展的决定者和使用者，在根本上人类是真正的责任者。“负责任”要贯穿人工智能的研发和应用，《新一代人工智能治理原则——发展负责任的人工智能》指出，要遵循“和谐友好、公平公正、包容共享、尊重隐私、安全可控、共担责任、开放协作、敏捷治理”的原则，这无疑为我们解决三重困境提供了治理框架和行动指南。“发展负责任的人工智能”，是人类以一种安全、可靠和合乎道德的方式研发和使用AI。“负责任”的内涵包括包容性、公平性、以人为本、透明和可解释性、安全性、尊重隐私和可被问责性等，以保障人工智能系统负责任地运行和使用。“负责任的人工智能”并不是要研发具有道德和责任意识的人工道德机器，使其能够进行自主思考和道德决策，并对自己的行为负责任。相反，“负责任的人工智能”主要指向的是人类要负责任，人类要肩负起对其所创造的人工智能的责任，负责任地发展人工智能使其更好地为人类服务。人类自身才是人工智能发展和治理的关键，因此我们必须要坚持以人为本的伦理立场。

Adhering to a people-centered ethical stance is the value foundation for promoting the development and governance of responsible AI. General Secretary Xi Jinping attaches great importance to the development, safety, and governance of AI, and for this purpose has proposed the important concept of “people-centered, AI for good.” [2]2 “People-centered, AI for good” (以人为本、智能向善) is a slogan introduced by Xi Jinping to guide AI ethics, emphasizing that AI development should prioritize human welfare and moral direction. This not only provides a new perspective for the development and governance of AI but also fully embodies the concept of a community with a shared future for mankind in the field of AI. People-centeredness contains deep ethical value and is an important principle to be followed in AI governance. The people-centered ethical stance requires that AI, in its research, development, and application, always adhere to the concept of “human beings as ends,” respecting human dignity and rights. As a technology, no matter how far AI develops, it should not shed its tool attribute; in essence, it is a product of human practical activity, i.e., an artifact. It is worth emphasizing that this people-centeredness is not a strong anthropocentric view that only highlights human dominance and superiority, but rather respects human personality and dignity, regards humans as ends, and emphasizes human responsibility and obligations.

坚持以人为本的伦理立场，是推动负责任人工智能发展和治理的价值基础。习近平总书记高度重视人工智能的发展、安全和治理，为此提出了“以人为本、智能向善”的重要理念。这既为人工智能的发展和治理提供了新视角，也是人类命运共同体理念在人工智能领域的充分体现。以人为本蕴含着深层的伦理价值，是人工智能治理要遵循的重要原则。以人为本的伦理立场，要求人工智能在研发和应用中要始终坚持“人是目的”的观念，尊重人的尊严和权益。人工智能作为技术，不管发展到何种程度，都不应摆脱工具属性，其在本质上是人的实践活动的产物，即人造物。值得强调的是这种以人为本并不是强人类中心主义观点，只突出人的主导性和优越性地位，而是尊重人的人格和尊严，把人看作目的，强调人的责任和义务。

While adhering to the people-centered ethical stance, to better develop responsible artificial intelligence, we must further establish the “principle of intelligent goodness,” the “principle of dignity,” and the “principle of responsibility” to respond to the threefold moral dilemma facing human-machine ethical relations. First, intelligent goodness is the fundamental goal of developing responsible AI. While AI brings great benefits to humanity, it also brings unprecedented challenges and risks. What we need is not technological neutrality, but to orient toward the purpose of goodness, to inject humanistic and ethical dimensions into technology and AI development, and to promote the advancement of AI in the direction of technology for good. This purpose of intelligent goodness cannot remain merely at the level of enhancing capabilities and satisfying the needs of the subject, but must aim to enhance the well-being of all humanity. Second, the principle of dignity is an inevitable requirement for developing responsible AI. Dignity is an important concept in ethics and a core element of the people-centered approach. The research, development, and application of AI are intended to better realize “the human being as an end,” embodying human dignity and value, rather than ultimately reducing humans to tools. Thus, the principle of dignity, as an inevitable requirement and ethical bottom line of the people-centered approach, demands that the development of AI must not harm or infringe upon human dignity and rights, must respect human autonomy, protect human privacy, and avoid algorithmic discrimination and intelligent harm. Finally, the principle of responsibility is an important guarantee for developing responsible AI. The principle of responsibility is one of the most critical principles in contemporary discussions of applied ethics and is also the most basic ethical principle that should be upheld in the research, development, and application of AI. The “people-centered” emphasis of the people-centered ethical stance is not only reflected in serving humanity and enhancing human well-being, but also in the fact that humans can and should be responsible, which is precisely the humanistic spirit of the people-centered approach. In the research, development, and application of AI, the principle of responsibility must always be upheld, integrating responsibility into every link of AI development, establishing and improving necessary accountability mechanisms, ensuring traceability of responsibility, and safeguarding the basic rights and interests of humanity.

在坚守以人为本的伦理立场下，为更好发展负责任的人工智能，我们还要进一步确立“智能向善原则”“尊严原则”和“责任原则”来回应人机伦理关系面临的三重道德困境。首先，智能向善是发展负责任人工智能的根本目标。人工智能给人类带来巨大福祉的同时，也带来了前所未有的挑战和风险。我们需要的并不是技术的中立，而是要朝向善的目的，在技术和人工智能的发展中注入人文和伦理维度，推动人工智能朝着科技向善的方向发展。这种智能向善的目的，不能仅仅停留在能力增强和主体需求的满足上，而是要以增进全人类的福祉为目标。其次，尊严原则是发展负责任人工智能的必然要求。尊严是伦理学的重要概念，也是以人为本的核心内容。人工智能的研发和应用是为了更好地实现“人是目的”，体现人的尊严和价值，而不是使人最后沦落为工具。由此，尊严原则作为以人为本的必然要求和伦理底线，要求人工智能的发展不能伤害和侵犯人的尊严和权益，要尊重人的自主性，保护人的隐私，避免算法歧视和智能伤害。最后，责任原则是发展负责任人工智能的重要保障。责任原则是当代应用伦理学讨论中最为关键的原则之一，也是人工智能研发和应用中应该坚持的最基本的伦理原则。以人为本的伦理立场所强调的“以人为中心”，不仅仅体现在为人类服务和增进人类福祉，更体现在人能够而且应该负责任上，这恰恰是以人为本的人文精神所在。在人工智能的研发和应用中，要始终坚持责任原则，将负责任贯穿到人工智能发展的各个环节中，建立健全必要的问责机制，做到责任的可追溯，保障人类的基本权益。

Building a Responsible AI Governance System

构建负责任的人工智能治理体系

In a life-world where humans and machines coexist, new ethical issues and moral dilemmas are gradually emerging. How to build a responsible AI governance system has become a key issue in AI development and the governance of science and technology ethics. We must adhere to the people-centered ethical stance, jointly build a responsible global AI governance system, form a mechanism of collaborative co-governance and shared responsibility, cultivate human AI literacy and a sense of responsibility, and jointly promote the progress of human civilization.

在人机共存的生活世界中，新的伦理问题和道德困境逐渐凸显出来，如何构建负责任的人工智能治理体系成为人工智能发展和科技伦理治理的关键问题。我们要坚持以人为本的伦理立场，共建负责任的全球人工智能治理体系，形成协同共治的责任机制，培育人类的人工智能素养和责任意识，共同推进人类文明的进步。

Building a responsible AI governance system requires jointly building a global AI governance system. Global AI governance is a common issue facing all countries in the world, concerning the fate of all humanity. In this community with a shared future for mankind in the new era of science and technology, all countries should participate in global cooperation and governance of AI with a responsible attitude, narrow the intelligence gap between countries, and jointly improve global governance mechanisms. The most important thing in building a responsible global AI governance system is to base it on the concept of a community with a shared future for mankind, aim to enhance the common well-being of all humanity, take the safety and rights of humanity as a prerequisite, and ensure that AI develops in a direction that advances human civilization. As a responsible major country in AI, China has always actively embraced intelligent transformation, vigorously promoted innovative AI development, and in the face of the opportunities and challenges AI brings to humanity, has insisted on being an active promoter, participant, and contributor in the field of global AI governance. China attaches great importance to AI safety governance and has implemented a series of practical measures, such as releasing the “Global AI Governance Initiative,” which emphasizes adhering to “people-centered” and “intelligent goodness,” providing the world with a Chinese solution for AI governance based on the concept of a community with a shared future for mankind.

构建负责任的人工智能治理体系，需要共建全球人工智能治理体系。人工智能全球治理是世界各国面临的共同课题，关乎全人类的命运，在这个新科技时代的人类命运共同体中，各国应以负责任的态度参与人工智能的全球合作和治理，缩小国家间的智能鸿沟，共同完善全球治理机制。构建起负责任的全球人工智能治理体系，最为重要的是立足人类命运共同体理念，以增进全人类共同福祉为目标，以保障人类安全和权益为前提，确保人工智能朝着人类文明进步的方向发展。我国作为负责任的人工智能大国，始终积极拥抱智能变革，大力推进人工智能创新发展，在面对人工智能给人类带来机遇和挑战的情况下，坚持做全球人工智能治理领域的积极推动者、参与者和贡献者。中国高度重视人工智能安全治理，实施了一系列务实举措，如发布《全球人工智能治理倡议》，强调坚持“以人为本”“智能向善”，为世界提供了基于人类命运共同体理念的人工智能治理中国方案。

Building a responsible AI governance system requires constructing a mechanism of collaborative co-governance and shared responsibility. A responsible agent refers not only to a single responsible subject but to an agent jointly constituted by multiple responsible subjects, which cannot be separated from the joint participation and collaborative co-governance of multiple subjects including the government, enterprises, industries, and users. The construction of a responsibility mechanism cannot simply attribute responsibility to AI developers and users; rather, it requires the participation of all responsible subjects, clarifying the subject, level, distribution, and boundaries of responsibility, and exploring a common responsibility mechanism to solve the responsibility gap and ethical vacuum. Although AI currently participates in human actions with relatively low autonomy and does not possess the moral subject status of humans, together with us, it constitutes human-machine interactive actions. Therefore, we must still incorporate it into a broader ethical framework and responsibility mechanism. Here, the responsible agent is a human-machine joint responsibility body composed of multiple responsible subjects and AI, requiring continuous improvement and refinement of the collaborative co-governance responsibility mechanism.

构建负责任的人工智能治理体系，需要建构协同共治的责任机制。负责任的责任者不仅是指单个的责任主体，而且是一个由多元责任主体共同构成的责任者，离不开政府、企业、行业、用户等多元主体的共同参与和协同共治。责任机制的构建，不能简单地将责任归结在人工智能研发者和使用者身上，而是需要所有的责任主体参与其中，需要明晰责任的主体、层次、分布和边界，探寻共同的责任机制，以解决责任鸿沟和伦理真空问题。虽然人工智能目前以比较低的自主性参与到人类行动中，并且不具有人类的道德主体地位，但它们与我们一同构成了人机交互行动。因此，我们仍要将其纳入更广泛的伦理框架和责任机制中。这里的责任者是一种由多元责任主体和人工智能共同构成的人机联合责任体，需要不断健全和完善协同共治的责任机制。

Building a responsible AI governance system requires cultivating human AI literacy and a sense of responsibility. Although constructing a global AI governance system and responsibility mechanism is very important for AI governance, we cannot rely solely on institutions and mechanisms; this is far from sufficient for building a responsible AI governance system. We must also attach importance to cultivating and enhancing the moral literacy and sense of responsibility of responsible agents, and we must implement the ethical identity of responsible agents onto every individual. As members living together in a community with a shared future for mankind, everyone must share the consensus that “everyone is a responsible agent.” In the era of AI, humanity not only needs to enhance AI scientific and technological literacy but also to cultivate AI moral literacy and a sense of responsibility. This cultivation cannot remain merely at the level of learning scientific and technological ethics knowledge; it must also attach importance to the cultivation of the heart, which is precisely what AI cannot truly replace in humans.

构建负责任的人工智能治理体系，需要培育人类的AI素养和责任意识。虽然构建全球人工智能治理体系和责任机制对于人工智能的治理非常重要，但我们不能仅仅依赖制度和机制，这对构建负责任的人工智能治理体系是远远不够的。我们还要重视培育和提升责任者的德性素养和责任意识，我们要将责任者的伦理身份落实到每一个人身上。作为共同生活在人类命运共同体中的成员，每个人都要有“人人都是责任者”的共识。在人工智能时代，人类不仅需要提升AI科技素养，还要培育AI德性素养和责任意识，这种培育不能仅停留在科技伦理知识的学习上，还要重视心的培育，这才是人工智能无法真正代替人类的东西。

The era of AI is a life-field where humans and machines coexist; AI has deeply intervened in our life-world. In this life-world of human-machine coexistence, we must adhere to the people-centered ethical stance, uphold the “principle of intelligent goodness,” the “principle of dignity,” and the “principle of responsibility,” and jointly build a responsible AI governance system. Although the future is full of risks and uncertainties, we must never forget that humans are the true responsible agents, always bearing the heavy responsibility and mission of developing responsible AI.

人工智能时代是一种人机共存的生活场域，人工智能已经深入介入我们的生活世界。在人机共在的生活世界中，我们要坚持以人为本的伦理立场，坚守“智能向善原则”“尊严原则”和“责任原则”，共同构建负责任的人工智能治理体系。尽管未来充满风险和不确定性，但永远不要忘记人类才是真正的责任者，始终肩负着发展负责任人工智能的重任与使命。

(Author: Liu Jing, Associate Professor, School of Philosophy, Department of Marxism, Northeast Normal University)

（作者：刘静，系东北师范大学马克思主义学部哲学院副教授）

Online Editor: Huihui

网络编辑：慧慧