AI-Driven Performance Feedback and Worker Persistence Under Income Uncertainty

Abstract

The proliferation of platform-mediated contingent work has made income uncertainty a defining feature of contemporary labor markets, while artificial intelligence increasingly mediates performance feedback in these environments. This paper presents a system-level examination of how AI-driven performance feedback interacts with worker persistence under conditions of income volatility. Drawing on interdisciplinary scholarship spanning computer science, organizational behavior, behavioral economics, and law, we analyze the sociotechnical architectures that algorithmic feedback systems inhabit, and the structural trade-offs they impose on governance, fairness, privacy, and sustainability. We argue that AI feedback functions as a coordinating mechanism that can either amplify or attenuate the motivational consequences of income variability, depending on its design transparency, temporal granularity, adaptivity, and embeddedness in broader platform governance regimes. The analysis reveals fundamental tensions between precision personalization and worker autonomy, between predictive optimization and interpretive agency, and between feedback-induced persistence and the erosion of long-term skill development. Using a multi-paradigm lens, the paper explores how different socio-infrastructural models yield divergent outcomes for persistence, inequality, and platform resilience. Policy implications related to openness mandates, algorithmic auditing, and cooperative data trusts are evaluated, offering a forward-looking synthesis for scholars, engineers, and regulators grappling with the transformation of effort regulation by AI in uncertain income settings.