Preprint: A Practical Ethics of Thinking: Human-in-the-Loop Responsibility and Authorship in Academic Work with AI

Abstract

This paper examines how generative AI [genAI] changes not only academic writing but the processes of thinking and knowledge creation. GenAI can serve as a productive space of resonance by challenging assumptions and introducing new perspectives, while its linguistic coherence can also create an Illusion of Evidence, in which plausibility is mistaken for justified knowledge. The paper argues that conventional Human-in-the-Loop approaches insufficiently capture researchers’ responsibility once AI becomes involved in the formation and evaluation of ideas. It therefore develops a Practical Ethics of Thinking and proposes Human-in-the-Loop-Thinking as a framework for preserving intellectual autonomy, epistemic vigilance, intellectual self-assertion, transparency, and non-delegable responsibility. Its central claim is that while AI can support the creation of knowledge, knowledge itself, human judgement, and the execution of responsibility are not delegable to a machine.