Overcoming the Incentive Collapse Paradox

arXiv stat.ML / 3/31/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper addresses the incentive collapse paradox in AI-assisted task delegation, where accuracy-based payment schemes can require unbounded payments to sustain positive human effort as AI improves.
It proposes a sentinel-auditing payment mechanism that guarantees a strictly positive, controllable human-effort level at finite cost, independent of AI accuracy.
Using this incentive-robust foundation, the authors introduce an incentive-aware active statistical inference framework that jointly optimizes auditing rate plus active sampling and budget allocation across tasks.
Experiments show better cost–error tradeoffs than standard approaches that use active learning or auditing alone.

Abstract

AI-assisted task delegation is increasingly common, yet human effort in such systems is costly and typically unobserved. Recent work by Bastani and Cachon (2025); Sambasivan et al. (2021) shows that accuracy-based payment schemes suffer from incentive collapse: as AI accuracy improves, sustaining positive human effort requires unbounded payments. We study this problem in a budget-constrained principal-agent framework with strategic human agents whose output accuracy depends on unobserved effort. We propose a sentinel-auditing payment mechanism that enforces a strictly positive and controllable level of human effort at finite cost, independent of AI accuracy. Building on this incentive-robust foundation, we develop an incentive-aware active statistical inference framework that jointly optimizes (i) the auditing rate and (ii) active sampling and budget allocation across tasks of varying difficulty to minimize the final statistical loss under a single budget. Experiments demonstrate improved cost-error tradeoffs relative to standard active learning and auditing-only baselines.