Learning from the Ethics of AI – A Research Proposal on Soft Law and Ethics of AI

Abstract

This contribution outlines a research proposal combining ethical guidelines on AI and a law-as-data approach. Building upon the definitions of soft law discussed in legal scholarship, it proposes a way of structuring the regulatory landscape on AI and of addressing the question of what is included in the “soft law of AI” today. By adopting a building-blocks approach (combining distinct definitional components of soft law), the paper shows that the state of current soft law on AI depends on which position on international law one defends. Concretely, the paper firstly offers a complete codebook for identifying the different types of soft law. Secondly, it applies this codebook as a proof-of-concept for the research proposal by analyzing 40+ ethical guidelines and by clustering preliminary results according to the actor enacting the guidelines and the legally relevant effects they could deploy. Four paradigmatic types of soft law emerge: statist and international organization soft law, process-oriented soft law, expertise-oriented soft law, and de facto relevant standards soft law. These results illustrate the contributions which are to be expected from a law-as-data research proposal.