摘要： 司法人工智能,尤其是大语言模型的诞生及其运用于司法实践的可能,看起来会消除司法裁判中法学方法论的运用空间。司法人工智能的基本原理在于根据历史裁判数据的统计规律预测(生成)当下的数据,但裁判数据的信息处理不等同于司法裁判本身,计算也不等同于推理或论证。从应然的角度看,只要坚守司法裁判作为说理活动的性质,法学方法论就不应丧失其运用的空间。从实然的角度看,在事实认定领域,机器算法因现实认知能力限制和评价余地应对乏力,无法恰当地从证据材料中建构出案件事实;在法律适用领域,机器算法无法替代创造性的法律解释及其诸多方法,也无法进行真正的漏洞填补和法律修正。但语料库语言学在法律解释中的应用,说明数字技术的确为法学方法论带来信息来源和具体运用样态的改变。未来的算法与方法,应当是相互增强的关系。

关键词: 司法人工智能, 机器算法, 法学方法论, 事实认定, 法律适用

Abstract: Judicial artificial intelligence, especially the birth of the Big Language Model and its possibility of application to judicial practice, seems to eliminate the space for legal methodology in the judicial decision.The fundamental principle of judicial artificial intelligence lies in the prediction (generation) of current data based on the statistical regularities of historical adjudication data, but the information processing of adjudication data is not equivalent to the judicial decision itself, nor is the calculation equivalent to reasoning or argumentation.From the perspective of “desirability”, as long as the nature of judicial decision-making as a kind of reason-giving activity is adhered to, legal methodology should not lose its space.From the perspective of “feasibility”, in the field of fact determination, on one hand, machine algorithms cannot properly construct the facts of the case from the evidence materials due to the limited real cognitive ability and non-response to evaluative space.On the other hand, in the field of legal application, machine algorithms cannot replace creative legal interpretation and its various methods, nor can they carry out real gap-filling and legal amendments.However, the application of corpus linguistics in legal interpretation shows that digital technology has indeed brought changes in the source of information and specific application patterns to legal methodology.In future, algorithms and methods should be mutually reinforced.

Key words: Judicial Artificial Intelligence, Machine Algorithms, Legal Methodology, Fact Determination, Legal Application