The Agency of Artificial Intelligence Systems

A Bibliographical Point of View

Abstract

A real-world case of the bibliographic field, that is the publication of Beta Writer’s book, Lithium-ion batteries. A machine-generated summary of current research, prompts a reflection on what it means, in formal terms, to be an author. This reflection is part of the larger framework of the discussion about the ethics of Artificial Intelligence, the ethics of Artificial Intelligence actions. This discussion implies the idea that Artificial Intelligence has complete agency i.e., can be assimilated to a fully rational and intentional subject; and that therefore its behaviors are unbounded from human subject responsibility. Instead, we want to argue that Artificial Intelligence does not have complete agency and that instead in Artificial Intelligence systems at least shared agency is expressed or even that Artificial Intelligence systems do not have agency because it resides in those who design and operate them. We argue about this not in the abstract, on principles, but from an examination of Beta Writer’s book.