初探目的解釋在法律人工智慧系統之運用可能

No Vehicles in the Park?: On AI’s Teleological Interpretation in Law

本文說明法律推理的計算機模型在設計上的可能困境。文章指出，在目的被給定的條件下，雖然可能運用CMLRs進行目的推論，但仍舊無法消除規範性的質疑，因為有限範圍之目的推論是對於未來可能結果之預測，因此，可能世界如何與目的之規範事態相符合是一個關鍵問題。此外，本文直視「目的如何具有規範性」的議題，並指出法律推理的不確定性。在實踐層面，則透過「目的」之概念分析為CMLRs提供哲學反省，釐清其運用界線以及邊界問題。

Legal purposes not only contribute to the meaning of legal norms but also help lawyers arrive at legal conclusions, and even provide jus-tificatory force. Engineers in Artificial Intelligence and Law also recog-nize the importance of teleological reasoning in law and aim at building a computational model capable of teleological interpretation, no matter in models of statutory reasoning or case-based reasoning. This paper investigates the problem of teleological interpretation in CMLRs. It points out that CMLRs are capable of operating teleological reasoning within a limited scope but incompetent to deal with normative issues such as why certain values shall be realized through a teleological inter-pretation. This is because CMLRs in a purposive interpretation function to predict future possible outcomes, but it is another (fundamental) question regarding how this presupposed possible world fits into the normative state of affairs of a legal purpose. This article makes two con-tributions. First, it directly focuses on the form and structure of purpos-ive interpretation providing analysis and reflections on its applicative problem in CMLRs. Second, it clarifies the relationship between a legal purpose and meaning of a legal norm. This clarification helps CMLRs to understand to what extent a model design of teleological interpreta-tion is acceptable.