人工智能伦理风险及其治理——基于算法审计制度的路径

Ethical Risk and Its Governance of Artificial Intelligence:Based on the Path of the Algorithm Audit System

基于算法内核,人工智能产生了诸如算法歧视、算法黑箱等一系列伦理问题。这是主观上人们理性失衡与过度追逐利益的恶果,是客观上算法技术缺陷与算法工程师水平局限的结果,也是法律规制落后与伦理教育缺位的后果。但现阶段算法技术的发展处于较为尴尬的境地:一方面,它已产生极大伦理隐患,具有治理的现实紧迫性;另一方面,它又处于技术上升期,过度约束可能阻碍产业发展。对此,算法审计制度以其适度约束力和有限透明的特点,可解决当前人工智能算法伦理治理的两难困境。

Due to the core of algorithms,artificial intelligence has generated a series of ethical issues such as algorithmic bias and black box.This is the results of human beings'rational imbalance and excessive pursuit of interests,the limitation of algorithm technology and engineers,as well as the outdated laws and the absence of ethical education.But now,the algorithm technology is in a relatively awkward position:on the one hand,it has generated great ethical hazards and needs to be governed;on the other hand,it is in the rising period,and excessive constraints may hinder its development.In this situa-tion,Algorithm Audit System,with its moderate force and limited transparency,can solve the dilemma of ethical governance of artificial intlligence.