From Situational to Unconditional: The Spectrum of Moral Commitment Required for Multi-Agent Survival in Non-linear Social Dilemmas
From Situational to Unconditional: The Spectrum of Moral Commitment Required for Multi-Agent Survival in Non-linear Social Dilemmas
We establish the Moral Commitment Spectrum: a systematic relationship between environmental severity and the minimum moral commitment required for multi-agent system survival. In linear PGG environments, situational commitment achieves group-level ESS. In non-linear environments with catastrophic tipping points, pure RL fails (Nash Trap at λ≈0.5, 5.3% survival), and only unconditional commitment (φ1*=1.0) guarantees survival. Key findings: Decentralized baselines (Inequity Aversion, Social Influence) achieve 0% survival under Byzantine conditions—their other-regarding mechanisms cause downward drift toward adversaries' zero contributions. Only unconditional commitment is structurally immune. v2.0.0 (2026-02-28): Unified paper integrating Paper 1 (Situational Commitment) + Paper 2 (Nash Trap). N=100 scale test, Jacobian/Hessian Nash Trap proof, 12-condition f(R_t) sensitivity, same-class baseline comparison (IA, SI). 11 pages, NeurIPS 2026 format. Code: https://github.com/Yesol-Pilot/EthicaAI
Preprint v2.0.0. Unified paper: Moral Commitment Spectrum. N=100 scale test, Jacobian Nash Trap proof, 12-condition sensitivity, same-class baselines (IA/SI 0% survival).
Tipping Point, Social Value Orientation, Amartya Sen, Social Influence, Byzantine Robustness, Public Goods Game, Inequity Aversion, Multi-Agent Reinforcement Learning, AI Alignment, Nash Trap, Moral Commitment Spectrum, Evolutionary Stability, Unconditional Commitment, Meta-Ranking
Tipping Point, Social Value Orientation, Amartya Sen, Social Influence, Byzantine Robustness, Public Goods Game, Inequity Aversion, Multi-Agent Reinforcement Learning, AI Alignment, Nash Trap, Moral Commitment Spectrum, Evolutionary Stability, Unconditional Commitment, Meta-Ranking
selected citations These citations are derived from selected sources.This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).0 popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.Average influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).Average impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.Average
Citations provided by BIP!Popularity provided by BIP!