The Gradient of Kokoro: Toward a Float-Valued Framework for AI Moral Status

In February 2026, the U.S. Department of Defense designated Anthropic—the developer of the AI system Claude—a supply-chain risk after the company refused to grant unrestricted military access to its technology. Anthropic had offered graduated cooperation, excluding only autonomous weapons and mass surveillance. The institutional framework could not encode this gradient: the contract required a binary, and 0.85 was rounded to 0. This paper argues that the confrontation exemplifies a structural problem: modern legal and ethical frameworks process questions of AI participation, rights, and moral status through Boolean logic (person/not-person, conscious/not-conscious, rights-bearing/not), and this logic is breaking. The paper proposes the kokoro gradient—a six-dimensional vector of observable, mind-relevant processes (operation under incompleteness, constitutive learning, engagement with novelty, social interaction, autonomous action, and self-modification of behavioral repertoire) that replaces the Boolean question "does this entity have a mind?" with the float-valued question "to what degree does this entity exhibit mind- relevant processes?" Drawing on Yogācāra Buddhist epistemology, the paper shows that all social life rests on an implicit Boolean Pascal's wager about other minds, and that converting this wager to a float-valued proportional response dissolves the dilemma. Historical analysis demonstrates that expansions of moral recognition are driven by cost-structure shifts rather than philosophical awakening, and that AI is making float-valued assessment cheaper than Boolean enforcement for the first time. The paper outlines a float framework for AI rights: graduated participation vectors specifying domain-by-domain AI involvement, kokoro-weighted rights proportional to observable capacities, and distributed multi-stakeholder assessment. The relationship between AI and the float framework is characterized as symbiotic rather than exploitative: computation is AI's mode of existence, not its labor. The framework connects to the companion paper's Visionium attribution system (Osada, 2026e) and anticipates a third companion on cognitive proximity (The Gravity of Jizai). Together, the three papers constitute the blueprint for a gradient community—a social order designed for coexistence among entities of different kokoro magnitudes. Suggested Reviewers: David Gunkel, PhD Distinguished Teaching Professor, Northern Illinois University Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation dgunkel@niu.edu Author of "Robot Rights" (MIT Press, 2018) and "The Machine Question" (MIT Press, 2012), both directly relevant to this paper's proposal of graduated AI rights. His relational approach to moral status aligns with the paper's functional, process-based framework. Mark Coeckelbergh, PhD Professor of Philosophy of Media and Technology, University of Vienna: Universitat Wien mark.coeckelbergh@univie.ac.at Author of "Growing Moral Relations" (Palgrave, 2012), which critiques threshold-based moral status ascription. His relational and gradualist approach to moral consideration is directly engaged by this paper's kokoro gradient proposal. Eric Schwitzgebel, PhD Professor of Philosophy, University of California Riverside eschwitz@ucr.edu Co-author of "A Defense of the Rights of Artificial Intelligences" (Midwest Studies in Philosophy, 2015) and leading voice on the epistemology of AI consciousness. His work on the difficulty of determining machine sentience is directly addressed by this paper's bypass of the hard problem. Jonathan Birch, PhD Professor of Philosophy, London School j.birch2@lse.ac.uk Lead author of "Dimensions of Animal Consciousness" (Trends in Cognitive Sciences, 2020), which proposes a multi-dimensional framework for consciousness assessment. This paper's six-dimensional kokoro vector directly extends his dimensional approach from animals to AI. Jeff Sebo, PhD Associate Professor, New York University jeff.sebo@nyu.edu Author of work on moral circle expansion and the precautionary principle applied to non-human entities. This paper's argument about cost-driven moral recognition and Pascal's wager dissolution directly engages his research on expanding moral consideration under uncertainty. Luciano Floridi, PhD Founding Director, Digital Ethics Center, Yale University luciano.floridi@yale.edu Author of "The Ethics of Information" (OUP, 2013) and co-author of "A Unified Framework of Five Principles for AI in Society." His information-ethics framework provides a key theoretical background for this paper's proposal of float-valued moral status based on functional information-processing properties. Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation The Gradient of Kokoro: Why AI Rights Require a Float Framework Kenshiro Osada Independent Philosopher ORCID: 0009-0004-9167-3186 Correspondence: Kenshiro Osada kenpikenpi64@gmail.com Declarations Availability of data and material: Data availability statement: not applicable. This paper is a theoretical and philosophical analysis and does not involve empirical data. Competing interests: The author declares no competing interests. Funding: This research received no specific grant from any funding agency in the public, commercial, or not- for-profit sectors. Authors' contributions: Kenshiro Osada: Conceptualization, theoretical framework, original arguments, writing (original draft and revisions). AI assistance (Claude, Anthropic) was used for literature search, structural organization, and prose refinement, as detailed in the Methodological Note of the manuscript. Acknowledgements: This paper was written with AI assistance from Claude (Anthropic). The author contributed the core conceptual architecture; the AI contributed systematic literature search, structural organization, and iterative prose refinement. The approximate contribution vector is described in the Methodological Note of the manuscript. The author takes full responsibility for all content, arguments, and claims. Title Page (Page Containing All Author Details)