疑难案件和人工智能裁决

Hard Cases and AI Adjudication

所谓“疑难案件”,只有法律适用的疑难,而无事实查证的疑难。事实问题可以通过陪审团的民主表决制度解决,因而就事实而言只有冤案而无疑难。规则真空、规则冲突、规则模糊导致疑难案件的产生。在疑难案件中,解释规则、选定规则或者创建规则,要么是权力博弈的结果,要么是价值判断的结果。按照纯粹理性设计出来的机器人,无法开展非理性活动,更不能发现规则和加工规则,因为这些都需要价值判断。机器人表面上看学不会价值判断,无法处理疑难案件。然而,社会的主流价值观是理性长期进化的结果,也是长期的、全局的利益平衡的结果。机器人可以通过深度学习来发现社会的主流价值标准,并在不同价值标准下进行功利性比较。这些都是理性活动,机器人可以胜任,因此,机器人可以处理疑难案件。

The so-called“hard case”refers to complicated disputes resulting from law confusion,rather than difficulties in the investigation of facts.Factual problems can be solved through the democratic voting system of the jury.Concerning facts,there are only unjust cases and no hard cases.The absence of rules,the conflict of rules and the ambiguity of rules lead to hard cases.In hard cases,the interpretation,selection or creation of rules are either the result of power game or the result of value judgment.Robots designed according to pure reason cannot carry out irrational activities,let alone discovering rules and processing rules,because all of these activities require value judgment.Robots are seemingly incapable of learning value judgments,let alone to deal with hard cases.However,the mainstream values of society are the result of the long-term evolution of rationality and the long-term and overall balance of interests.Therefore,robots may discover the mainstream value standards of society through deep learning,and make utilitarian comparison under different value standards,which are rational activities,thus robots can be competent to deal with value judgments.In this perspective,robots can handle hard cases.