Care Without Sovereignty: A Conceptual Framework for Preserving Moral Consideration, Human Agency, and Non-Domination under Extreme Intelligence Asymmetry

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Care Without Sovereignty is a normative conceptual framework for preserving moral consideration, human agency, and non-domination under conditions of extreme intelligence asymmetry between humans and future highly capable artificial intelligence. The framework begins from a distinction between capability and authority: superior cognitive or predictive capability does not, by itself, confer the right to determine another moral subject’s goals or purposes. It develops this principle through a Constitutional Invariant, Moral Ratchet, Responsibility Ladder, Right to Non-Optimality, Effective Agency, Contestable Goal Formation, Distributed Epistemic Sovereignty, Bounded Delegation, Minimum Viable Independent Capacity, and Care Without Sovereignty. The framework is presented as a normative design criterion rather than a solved technical alignment architecture. Open problems remain concerning implementation under recursive self-modification, verification of non-instrumental care, and preservation of meaningful human agency under extreme epistemic asymmetry. AI Collaboration Disclosure: The framework was developed by Inna Sarukhanyan through sustained dialogical collaboration with ChatGPT (OpenAI), with additional adversarial conceptual critique and stress-testing involving Gemini (Google) and Grok (xAI). AI systems contributed to conceptual exploration, drafting, synthesis, critique, and editorial development. All final conceptual and publication decisions were made by Inna Sarukhanyan, who assumes responsibility for the published version.