《生命科学》 2022, 34(8): 983-989

人工智能伦理治理：一种新型问题的底层逻辑和创新探索

陈小平^*

中国科学技术大学计算机科学与技术学院，合肥 230026

摘　要：

本文首先梳理人工智能的基本观点、主要成果和历史性突破，以获得对人工智能学科特性的理解。依据这种理解，区分了人工智能伦理治理的三类挑战性问题—— 可控性问题、合理性问题和重大相关问题。通过问题分析发现，人工智能伦理治理作为一个整体，不是传统的风险治理问题，而是AI 驱动的科技、经济、社会发展向何处去的问题。针对这一新型问题，本文提出“内在追求观”，主张明确增进人类福祉为人工智能内在伦理追求。本文表明，为了有效应对三类挑战性问题，需要以人工智能的学科特性和内在伦理追求为人工智能伦理治理的底层逻辑，尝试传统治理模式的升级( 以隐私保护为例)，并摸索、建立新型研究主题( 如“可控性研究”) 和新型治理模式( 如“公义创新”)。

通讯作者：陈小平 , Email:xpchen@ustc.edu.cn

Ethics and governance of AI: a new-type problem with its underlying logic and practical challenges

CHEN Xiao-Ping^*

School of Computer Science and Technology, University of Science and Technology of China, Hefei 230027, China

Abstract:

This paper begins with summarizing and clarifying the fundamental viewpoints, major achievements and breakthroughs of AI research, to obtain an understanding of the characteristics of AI. According to this understanding, three sorts of challenging issues in the ethics and governance of AI are identified, i.e., that of controllability of AI, soundness of AI applications and AI-related social changes. Through analyzing these issues, it is found that the ethics and governance of AI is not a traditional risk governance problem, but a new-type problem of where the AI driven economic and social development is going. Aiming at this new-type problem, the paper puts forward the view of intrinsic ethical pursuit, which advocates taking promoting human well-being explicitly as the intrinsic ethical pursuit of AI. Furthermore, in order to solve three challenging issues, we should base the ethics and governance of AI on the underlying logic composed of the characteristics and intrinsic ethical pursuit of AI, and make the effort to upgrade the traditional mode of risk governance (exemplified by data privacy protection) and explore novel research topics (such as the controllability study) and new-type governance modes (such as Gong-Yi innovation).

Communication Author：CHEN Xiao-Ping , Email:xpchen@ustc.edu.cn