作者：刘东亮
作者介绍：西安交通大学法学院教授
出版日期：2022年6月9日
内容提要 法律智能系统运作逻辑的合理性，是取得当事人和公众信任的基础，也是其决定具有正当性和可接受性的重要根据。传统的基于规则的专家系统，其逻辑根基是演绎推理。基于案例的推理的专家系统，底层逻辑是类比推理。使用大数据技术的智能系统，属于“基于计算模型的推理”。在处理广泛存在的模糊性问题时，智能系统会运用到模糊推理和模糊逻辑。由于新一代智能系统多采用混合技术路径，上述推理方式可以并存。鉴于价值选择和价值判断是现阶段计算机能力的短板，智能系统尚无法像人类一样展开情理法兼容的论证，解决出路是将论证的复杂要求“降维”成解释，即非正式的说明理由。可以合理预见，法律智能系统的发展方向是“人机协同”。
关键词 法律智能系统 逻辑推理 论证说理 人机协同
Abstract: The rationality of the operation logic of the legal intelligence systems is the basis for obtaining the trust of the parties and the public, and also an important basis for the legitimacy and acceptability of their decisions. The logical basis of the traditional rule-based expert system is deductive reasoning. The underlying logic of the case-based reasoning expert system is analogical reasoning. The intelligent systems using big data technology can be attributed to ‘reasoning based on computational models’. In dealing with widespread fuzzy problems, intelligent systems are applied to fuzzy reasoning and fuzzy logic. Since the new generation of intelligent systems mostly adopt hybrid technology paths, the above reasoning methods can coexist simultaneously. In view of the fact that value selection and value judgment are the weakness of computer ability at the present stage, intelligent systems cannot develop reasonable and compatible arguments like human beings, and the solution is to lower the complex requirements of arguments to informal interpretation. It can be reasonably predicted that the development direction of legal intelligence system is ‘man-machine collaboration’