Implicit decision voting made by humans as normative and implementable rules with the help of language models

Patrick Hedfeld ; Rüdiger Buchkremer/Oliver Koch/Andreas Lischka (Hrsg.)

Can an ethically justifiable decision-making process be facilitated through a machine moral agent in the form of advice? The premise revolves around decision votes explicitly solicited from human individuals and based on scenarios such as the Trolley Problem, reflected in data and processed through generative language models. These advisories can then be formulated in a general manner and discussed within a societal context, emerging implicitly from individual decisions. Furthermore, we discuss the concept of an implicit moral agent and an honorable AI advisor.

MoreLess

Year of publication:	2025
Authors:	Hedfeld, Patrick
Other Persons:	Buchkremer, Rüdiger (ed.) ; Koch, Oliver (ed.) ; Lischka, Andreas (ed.)
Publisher:	Essen : MA, Akademie Verlags- und Druck-Gesellschaft mbH
Subject:	AI \| Generative Language Models \| Decision Votes \| Trolley Problem \| Implicit Agent \| Theorie \| Theory \| Sprache \| Language \| Entscheidung \| Decision \| Abstimmungsregel \| Voting rule \| Künstliche Intelligenz \| Artificial intelligence \| Neue politische Ökonomie \| Public choice