LLM-policy Provisional version

This repository contains a preprint version of the manuscript “Pre-Ontological Framework for AI Agency”. The paper proposes a conceptual framework that distinguishes among thinking, intelligence, consciousness, emotion, being alive, and human life, prior to ontological or normative commitments. It argues that many contemporary confusions surrounding AI agency, responsibility, and regulation arise from insufficient conceptual differentiation rather than from empirical uncertainty. Within this framework, large language models are analyzed as systems that instantiate forms of thinking and limited intelligence, while lacking consciousness and intrinsic agency. Apparent life-like or agent-like intuitions are explained in terms of relational irreversibility and socio-technical embedding, rather than subjectivity. This manuscript is a preprint and has been submitted for peer review to Philosophy & Technology (Springer Nature).The content may be revised in response to editorial and reviewer feedback.