🤖 AI Summary
This work addresses decision failure in high-stakes intelligent agent systems, arguing that it stems primarily from excessive upstream concentration of choice rights rather than merely misaligned objectives. The authors propose an incentive-based governance mechanism that models choice rights as a constrained reinforcement learning process, projecting policy updates onto a governance-defined feasible set at each iteration to ensure bounded discretion. By integrating learnable scoring and shrinkage parameters with a quantification of governance debt, the approach uniquely enables adaptive optimization under sovereignty constraints while simultaneously decentralizing choice rights in dynamic environments. Empirical evaluations across multiple financial regulatory scenarios demonstrate that the method effectively prevents the emergence of deterministic monopolies induced by unconstrained reinforcement learning, achieving sustained performance improvements within bounded choice regimes.
📝 Abstract
Selection as Power argued that upstream selection authority, rather than internal objective misalignment, constitutes a primary source of risk in high-stakes agentic systems. However, the original framework was static: governance constraints bounded selection power but did not adapt over time. In this work, we extend the framework to dynamic settings by introducing incentivized selection governance, where reinforcement updates are applied to scoring and reducer parameters under externally enforced sovereignty constraints.
We formalize selection as a constrained reinforcement process in which parameter updates are projected onto governance-defined feasible sets, preventing concentration beyond prescribed bounds. Across multiple regulated financial scenarios, unconstrained reinforcement consistently collapses into deterministic dominance under repeated feedback, especially at higher learning rates. In contrast, incentivized governance enables adaptive improvement while maintaining bounded selection concentration.
Projection-based constraints transform reinforcement from irreversible lock-in into controlled adaptation, with governance debt quantifying the tension between optimization pressure and authority bounds. These results demonstrate that learning dynamics can coexist with structural diversity when sovereignty constraints are enforced at every update step, offering a principled approach to integrating reinforcement into high-stakes agentic systems without surrendering bounded selection authority.