Internal vs. External: Comparing Deliberation and Evolution for Multi-Agent Constitutional Design
This work addresses the unresolved question of whether multi-agent AI system rules should emerge internally or be discovered externally, providing the first controlled comparison with concrete results.
The paper compares internal deliberation vs. external evolution for designing behavioral constitutions in multi-agent systems across three social environments. Evolution significantly outperforms deliberation in collective-action settings (p < 0.01), but neither method improves bilateral trading; evolution's advantage inverts under certain incentive shifts.
Multi-agent AI systems need behavioral constitutions, but it is unresolved whether such rules should emerge internally through agent self-governance or be discovered externally through optimization. We present the first controlled comparison of internal deliberation and external evolution across three social environments: a coordination grid-world, an iterated public goods game, and a bilateral trading market. Across 180 simulation runs, evolution significantly outperforms deliberation in collective-action settings (p < 0.01), while neither method improves outcomes in bilateral trading. A multiplier ablation reveals that evolution's advantage inverts when incentives shift: at pool multiplier (m = 0.75) the evolved constitution forces value-destroying cooperation and becomes the worst-performing method. Notably, no deliberation run across thirty trials ever proposed punishment -- the canonical cooperation-sustaining mechanism evolution reliably discovers -- suggesting external optimization wins on peaks while internal self-governance trades peaks for structural responsiveness.