人工道德智能体何以可行——基于对价值敏感性设计的审视

How to Feasibly Develop Artificial Moral Agents:Based on the Review of Value Sensitive Design

人工道德智能体自主程度越来越高,其引发伦理困境的实质是自主机器系统的"技术偏见"与人类伦理价值的冲突。当前人工道德智能体的建构依赖的特定技术模型或伦理理论并不完善。价值敏感性设计为人工道德智能体的实现提供多维视角,是一种整体性、动态性的实践方法论。在该框架下,人类价值被系统性分析并被嵌入到人工道德智能体之中,但该方法仍待完善,以实现人工道德智能体与人类社会的合理互构。

The Artificial Moral Agents are getting more and more autonomous.They often cause ethical dilemma,and the essence of the dilemma is the conflict between the"technical bias"of autonomous machines and the ethical values of human beings.The current construction methods of Artificial Moral Agents depend on specific technical models or ethical theories which are defective.Value sensitive design provides a holistic and dynamic practical methodology with multi-dimensional perspectives for the development of Artificial Moral Agents.With value sensitive design,the human values can be analyzed and embedded in the Artificial Moral Agents.The methodology still needs to be improved to promote the mutual shaping of Artificial Moral Agents and human society.