The AI Penalization Effect: People Reduce Compensation for Workers Who Use AI

2025-01-22

AI Summary

This study investigates whether AI tool usage affects workers’ compensation decisions. Through 11 controlled experiments (N = 3,846) spanning hypothetical and real-world gig settings, diverse job types, compensation schemes, and worker identities, we identify a robust “AI penalty effect”: employers systematically reduce pay for workers who use AI tools. This effect stems from an attribution bias—specifically, underestimating individual effort and contribution when AI is involved—and intensifies under weaker contractual constraints. Employing behavioral experiments, multimodal compensation measurement, mediation analysis, and a field experiment with real monetary payments, we robustly replicate the effect, isolate its core psychological mechanism, and demonstrate that greater contractual rigidity mitigates the resulting inequity. To our knowledge, this is the first study to name, empirically document, and causally unpack how AI adoption may exacerbate labor income inequality—providing critical evidence for safeguarding worker rights in the AI era.

Technology Category

• Philosophy and Ethics of AI: AI & Jobs/Labor
• Humans and AI: AI for Accessibility
• Machine Learning: Ethics, Bias, and Fairness

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

We investigate whether and why people might adjust compensation for workers who use AI tools. Across 11 studies (N = 3,846), participants consistently lowered compensation for AI-assisted workers compared to those who were unassisted. This"AI Penalization"effect was robust across (1) different types of work (e.g., specific tasks or general work scenarios) and worker statuses (e.g., full-time, part-time, or freelance), (2) different forms of compensation (e.g., required payments or optional bonuses) and their timing, (3) various methods of eliciting compensation (e.g., slider scale, multiple choice, and numeric entry), and (4) conditions where workers' output quality was held constant, subject to varying inferences, or statistically controlled. Moreover, the effect emerged not only in hypothetical compensation scenarios (Studies 1-9) but also with real gig workers and real monetary compensation (Studies 10 and 11). People reduced compensation for workers using AI because they believed these workers deserved less credit than those who did not use AI (Studies 7 and 8). This mediated effect attenuated when it was less permissible to reduce worker compensation, such as when employment contracts provide stricter constraints (Study 8). Our findings suggest that adoption of AI tools in the workplace may exacerbate inequality among workers, as those protected by structured contracts are less vulnerable to compensation reductions, while those without such protections are at greater risk of financial penalties for using AI.

Problem

• People reduce pay for workers using AI tools
• AI Penalization persists across diverse work scenarios
• AI use may worsen workplace inequality

Innovation

• Studied AI impact on worker compensation
• Tested across diverse work scenarios
• Explored mediation by perceived worker credit