人工智能机器伦理治理的进路主要有“外在主义”和“内在主义”2种，“外在主义”强调以人类制定的伦理准则规约人工智能机器的发展，而“内在主义”则突出人工智能机器的自学习能力，以形成独立于人类的机器主体道德意识。通过介绍人工智能机器伦理与属人的伦理的本质区别，揭示了属人的伦理世界包含着的审美性意涵，即通往道德意识的美与美感无法被模拟与分层，形成道德意识的有机生命与审美判断力无法被还原，从而推理出人工智能机器的主体道德意识无法产生，得出在功利主义原则驱动下的人工智能机器伦理不能通往人类伦理意义世界的结论。人工智能的道德意向性是形式逻辑推演的结果，人工智能机器缺乏“具身”，不能从事感性实践活动并产生情感意识，其认知模式属于单一的还原论。因此，人工智能机器不能自主生成美善价值，故其治理的“内在进路”也是“外在主义”进路的延伸与扩展。
There are two approaches to the ethical governance of AI machines: "externalism" and "internalism". "Externalism" emphasizes the development of AI machines based on the ethical norms formulated by humans, while "internalism" highlights the self-learning ability of AI machines to form the moral consciousness of machine subjects independent of humans. By introducing the essential differences between artificial intelligence machine ethics and human ethics, this paper reveals the aesthetic implications contained in the human ethical world, that is, the beauty and aesthetic sense leading to moral consciousness cannot be simulated and stratified, and the organic life and aesthetic judgment forming moral consciousness cannot be restored. The subject moral consciousness of AI machine cannot be produced and the AI machine ethics driven by the principle of utilitarianism cannot lead to the conclusion of the meaning world of human ethics. The moral intentionality of artificial intelligence is the result of formal logic deduction, artificial intelligence machine lacks "embodiment", cannot engage in perceptual practice activities nor produce emotional consciousness, and its cognitive mode belongs to a single reductionism. Therefore, artificial intelligence machines cannot generate beauty value independently, so the "internal approach" of its governance is also an extension and expansion of the "externalism" approach.
[1] 陈肖东.预见人工智能与人的智能之鸿沟——基于马克思认识论对认知神经科学的解构[J].理论界, 2021(9):22-28.
[2] 张卫.人工智能伦理的两种进路及其关系[J].云南社会科学, 2021(5):21-27.
[3] Moor J H. The nature, importance, and difficulty of machine ethics[J]. IEEE Intelligent Systems, 2006, 21(4):18-21.
[4] 庞洁.机器意识研究的演进道路及其意向性分析[J].网络安全和信息化, 2023(1):14-17.
[5] 亚里士多德.尼各马可伦理学[M].廖申白,译.北京:商务印书馆, 2003:18-19, 52-57.
[6] Pasquali A, Timmermans B, Cleeremans A. Know thyself:Metacognitive networks and measures of consciousness[J]. Cognition, 2010, 117(2):182-190.
[7] Fekete T, Edelman S. Towards a computational theory of experience[J]. Consciousness and Cognition, 2011, 20(3):807-827.
[8] Kitamura T, Tahara T, Asami I. How can a robot have consciousness?[J]. Advanced Robotics, 2000, 14(4):263-275.
[9] 陶艳华.马克思政治伦理思想研究[M].北京:人民出版社, 2009:5.
[10] 赵泽林.内部区分:意识与机器意识的新进展及其哲学前瞻[J].哲学动态, 2021(8):118-127.
[11] 赵汀阳.人工智能的自我意识何以可能[J].自然辩证法通讯, 2019(1):1-8.
[12] 刘晓男.康德审美道德论与浪漫主义思潮[J].学术交流, 2018(7):37-42.
[13] 马克思. 1844年经济学哲学手稿[M].北京:人民出版社, 2000:88.
[14] 赫伯特·马尔库塞.审美之维[M].李小兵,译.桂林:广西师范大学出版社, 2001.
[15] Gamez D. Human and machine consciousness[M]. Cambridge:Open Book Publishers, 2018.
[16] 岩城见一.感性论——为了被开放的经验理论[M].北京:商务印书馆, 2008:149.
[17] 马克思恩格斯选集(第4卷)[M].北京:人民出版社, 1995:936.
[18] 马克思恩格斯选集(第3卷)[M].北京:人民出版社, 2019:923.
[19] Henry M. Philosophy and phenomenology of the body[M]. The Hague:Springer, 1975:7-8.
[20] 蓝江.一般数据、虚体与数字资本——历史唯物主义视域下的数字资本主义批判[M].南京:江苏人民出版社, 2022:99.
[21] 马克思恩格斯全集(第2卷)[M].北京:人民出版社, 1957:152.
[22] 皮埃尔·布迪厄.实践感[M].蒋梓骅,译.南京:译林出版社, 2003:101.
[23] 弗洛姆.追寻自我[M].苏娜,安定,译.延吉:延边大学出版社, 1987:100.
