Non-Zero Ethics: A Risk-First Principle for Cooperation and Capability in Human–AI Societies

Non-Zero Ethics (NZE) offers a compact ordering of reasons for high-stakes choice: first reject options that raise the risk of civilizational extinction (including irreversible value lock-in); among the remainder, prefer those that expand widely shareable capabilities and stabilize cooperation, while protecting excellence through ring-fenced audacity. The view is constructivist-pragmatist and pluralist. NZE is justified via (i) a contractualist argument from mutual vulnerability and the preconditions of public reason, and (ii) a game-stability argument showing when cooperation outcompetes predation. An epistemic note—Hybrid Structural Humanism (HSH)—specifies identifiability and public-reason standards for claims about risk, capability, and stability. Entropy/information serve as constraints and diagnostics, not as values. Companion papers: Agents and Patients in Plural Societies (APPS) and A Teleology of Flourishing for Plural Societies. Preprint v1; comments welcome. A revised version will be submitted to [venue TBD]