A typology of automatically processable regulation

Abstract

The possibility of encoding regulation to make it processable automatically by computers has been gaining attention within the legal discipline. With it, an abundance of terms has emerged as much as an array of academic discussions providing different examples, raising different concerns, while, unfortunately, having different premises in mind. This makes contributions within the field of what we refer to as 'automatically processable regulation' difficult to compare with each other and research results hard to transfer among different research projects and groups. To overcome this problem, we propose a typology that enables researchers to locate their research project within the domain of automatically processable regulation, understand what issues might arise depending on where within the typology a project falls, and determine the relationship between projects. The typology revolves around three dimensions: the primary aim of the project, the potential for divergence of interests amongst stakeholders, and the degree of mediation by computers.

Algorithms in the court: does it matter which part of the judicial decision-making is automated?

“…Much like the others, this attempt aims to describe current technological solutions rather than provide a clear taxonomy. More recently, Whalen ( 2022 ) proposed to classify legal technologies according to their “legal directness” and “legal specificity”, 14 whereas Guitton et al ( 2022a , b ) suggest mapping regulatory technologies along three different dimensions 15 : the project’s aim, divergence of interests between stakeholders, and the degree of human mediation. A different approach was taken by Tamò-Larrieux et al ( 2022 ), who propose the concept of Machine Capacity of Judgment (MCOJ).…”