人工道德顾问和苏格拉底式虚拟助手--国外对道德人工智能研究的两种进路  

Artificial Moral Advisor and SocrAl:Two Approaches to Research on Moral Artificial Intelligence Abroad

作　　者：朱国义 吴瑾菁 ZHU Guoyi;WU Jinjing(School of Marxism,Jiangxi Normal University,Nanchang,Jiangxi,330022)

机构地区：[1]江西师范大学马克思主义学院,江西南昌330022

出　　处：《自然辩证法通讯》2024年第10期33-41,共9页Journal of Dialectics of Nature

基　　金：教育部人文社会科学项目思政课教研专项“中国共产党百年奋斗精神贯穿思政课教学研究”(项目编号:21SZK-10704001);江西省教育厅研究生创新基金项目“习近平生态文明思想的时代性贡献研究”(项目编号:YC2023-B116)。

摘　　要：道德人工智能的出场,初衷是为避开生物保守主义对生物医学道德增强(BME)的反对。随着研究深入,学者们发现人工道德顾问(AMA)虽能有效避免生物保守主义对生物医学道德增强干涉人的自由的反对意见,却也陷入影响人的自主性的窠臼。人工道德顾问的道德增强方案致力于实现,以高度发达的道德人工智能帮助人们合乎道德地决策。完美道德机器容易使人们产生道德行动的依赖性,人的道德生活有被道德机器宰制的风险。针对于此,苏格拉底式虚拟助手(SocrAI)继而被提出,将人工智能的定位由道德顾问转换为道德教育工具,以道德对话形式帮助人们合乎道德地推理,实现道德增强。The original intention of the proposal of moral artificial intelligence was to avoid the opposition of biological conservatism to Biomedical Moral Enhancement(BME).With the deepening of research,scholars have found that although Artificial Moral Advisors(AMA)can effectively avoid the opposition of biological conservatism to the interference of Biomedical Moral Enhancement in human freedom,they also fall into the trap of affecting human autonomy.The Ethics Enhancement Program of Artificial Moral Advisor is committed to achieving highly developed ethical artificial intelligence to help people make ethical decisions.A perfect moral machine can easily lead to dependence on moral actions,and there is a risk of human moral life being dominated by moral machines.In response to this,SocrAI was proposed to transform the positioning of artificial intelligence from a moral advisor to a tool for moral education,helping people to reason morally and enhance their moral abilities through moral dialogue.

关 键 词：非道德增强 人工道德顾问 苏格拉底式虚拟助手 人工智能 

分 类 号：B82[哲学宗教—伦理学] TP18[自动化与计算机技术—控制理论与控制工程]