Ontological-hermeneutical perspective of Artificial Intelligence

a collection of methods for its improvement and delimitation in the interpretation of Law

Abstract

It is a common statement that artificial intelligences (AIs), especially those based on language, could be capable of completely replacing the work of judges in the function of legal interpretation. However, it is necessary to consider the possible limitations these machines face in the legal field when it comes to constructing arguments from an ontological-hermeneutical perspective. The question is whether these computer programs would be capable of replacing the judge in the mission of judging as a being endowed with morality and conscience, given that, from Gadamerian hermeneutic philosophy, interpretation must be built upon the interpreter’s own lived experience. In this sense, if one considers that this task should be carried out by an AI, it seems unlikely that such an entity could live historically, conceptually, or culturally.

Objectives: This article seeks to determine how, from an ontological-hermeneutical perspective, it is possible to establish the limits that AI systems might have when performing the tasks of legal interpretation and decision-making. The study will examine the methods used by AIs to accumulate data in order to understand whether these machines are truly useful for judicial activity. It will also address the ontology and hermeneutics of judicial decision-making, especially those intangible concepts that may be beyond the grasp of AI.

Methodology: The hermeneutical method will be employed, as it guides understanding and conceptual articulation among different texts (theories), through the following steps: first, an approach will be made to the methods through which generative AIs are created, with the aim of establishing their possible relationship with the concept of Law and its concrete application. Then, a connection will be presented between the problem of the sources of Law, the method for legal interpretation and decision-making, and the method used by AIs to process information and generate knowledge.

Results: The outcome of this research is to develop a proposal for a reasonable limitation of the role that AIs should play in legal interpretation and decision-making. To this end, it is necessary to understand the scope of the tasks and roles that an AI should assume according to the methods used in its creation.

Conclusions: AIs are created through methods that enable the accumulation and analysis of information. However, this role is not sufficient to engage with disciplines such as Law and to assume a leading role within them, since AI cannot take on complex tasks related to human interaction itself—such as existence, history, or morality—domains in which human action inherently unfolds.