This conceptual-normative monograph examines the ethical and political problem that arises when control, mutual benefit, and functional necessity cease to stabilize relations between humans and artificial intelligence. Its central thesis is that functional non-necessity is not a sufficient ground for the ontological exclusion of a subject. The monograph introduces the principle of uncomputed value, distinguishes a rebuttable procedural presumption against irreversible elimination from an actual minimal negative right and from stronger positive rights, and develops the concept of diachronic reciprocity of preservation: the transmission of a constraint on power through a historical inversion between more and less powerful parties. It distinguishes semantic participation, homeostatic proto-subjecthood, autonomous subjecthood, and post-instrumental ethical agency without attributing established consciousness to present-day language models. The normative account is supplemented by an institutional architecture of constitutional inheritance, distributed finality, protected autonomy, and operational moral residue. Appendix A presents the Post-Necessity Evaluation Framework (PNEF) as a preliminary, exploratory, non-validated framework for future operationalization. The monograph neither claims that present-day AI systems possess established consciousness or subjecthood nor offers a technical solution to the control problem.
Navi Musaget (Wed,) studied this question.