算法裁判的理论建构及路径选择——基于若干人工智能实验的启示

The Theoretical Construction and Path Selection of Algorithmic Judgment——Enlightenment through Artificial Intelligence Experiments

智能裁判是智能社会治理的重要组成部分。然而,对于人工智能是否适用于司法裁判以及能否进一步强化公平正义,还存在强化说与弱化说二种学说。为了准确把握人工智能应用于司法裁判的困境根源及应对策略,实验团组进行了系列人工智能应用于证据审查、法律事实认定以及法律适用的试验,并以此为基础建构智能裁判的理论体系,包括计算正义理论、机器裁判权理论和算法可辩论理论;其中,计算正义理论强调算法对于推动司法裁判正义具有不可替代的价值,机器裁判权理论则强调以法官为主导、以机器为辅助的裁判格局,算法可辩论理论则强调算法的可辩论性为当事人的权利救济提供有效保障。以智能裁判理论为基础,寻求裁判未来发展的路径:从裁判数字化向裁判智能化转变系总体方向;寻找法官与算法模型的黄金分割点,实现人工智能精准应用系核心所在;构建AI裁判模型算法审查制度则成为智能裁判体系建设的根本保障。

Judgment is an important part of intelligent social governance.There are two theories about whether artificial intelligence is applicable to judicial judgment and whether it can further strengthen fairness and justice.To grasp the causes of the problem and find solutions,we conducted a series of experiments to evidence review,legal fact determination,and legal application.The theoretical system of intelligent adjudication has been constructed;it includes the theory of computational justice,the theory of machine jurisdiction,and the theory of algorithmic arguability.The theory of computational justice emphasizes the value of algorithms in promoting judicial justice.The theory of machine jurisdiction emphasizes that judgment shall be controlled by human judges but assisted by machines.The theory of algorithmic arguability focuses on the review and questioning of the algorithms in use,which provides effective protection for the rights of the parties.This paper describes the future of intelligent judgment:its transformation from digital to intelligent,the collaboration between judges and algorithmic models,and the construction of an algorithmic model review system.