NBER Working Paper Series
The concept of pro-worker AI challenges the assumption that artificial intelligence inevitably displaces labor. In Building Pro-Worker Artificial Intelligence, the authors argue that technological outcomes are not predetermined; rather, they depend on how innovation is designed, deployed, and regulated. Instead of accepting automation as a force that weakens workers, the paper proposes an alternative path in which AI complements human skills and strengthens bargaining power.
The working paper situates artificial intelligence within a broader historical debate about technology and labor markets. While past technological revolutions often increased productivity, they also reshaped wage structures and skill demands. The authors warn that without intentional design and policy intervention, AI may deepen inequality and reduce worker leverage.
The Problem: Automation Bias and Labor Displacement
A central concern of the paper is what might be called an automation bias—the tendency of firms to adopt technologies that replace workers rather than augment them. Economic incentives, tax structures, and competitive pressures often favor labor-saving innovations. As a result, even when AI could complement human tasks, market dynamics may push firms toward substitution.
The authors explain that this pattern has consequences beyond employment levels. Labor-displacing technologies can weaken worker bargaining power, suppress wage growth, and increase income concentration. Therefore, the direction of technological change becomes a political and institutional question, not merely a technical one.
Defining Pro-Worker AI
The authors define pro-worker AI as systems intentionally designed to enhance worker productivity, decision-making, and skill development rather than eliminate roles. Such systems assist workers by improving information access, supporting task coordination, and enabling higher-value contributions.
Importantly, the paper emphasizes that AI design choices are shaped by policy environments and organizational incentives. Whether artificial intelligence complements or replaces labor depends on regulatory frameworks, corporate governance structures, and labor market institutions.
In this sense, technology is not neutral. It reflects economic incentives and power relationships embedded within markets and institutions.
Policy Mechanisms and Institutional Design
To promote pro-worker AI, the authors propose a range of policy tools. These include tax reforms that reduce incentives for labor displacement, stronger labor protections, investment in training and reskilling programs, and public funding for complementary technologies.
Education policy also plays a key role. Preparing workers to interact effectively with AI systems increases the likelihood that technology will augment rather than substitute their skills. Moreover, collective bargaining structures and worker voice mechanisms can influence how firms deploy automation internally.
The paper argues that governments, firms, and labor institutions must coordinate to steer innovation toward shared prosperity rather than concentrated gains.
Broader Economic Implications
The long-term stakes of pro-worker AI extend beyond employment. The authors link the direction of technological change to productivity growth, income distribution, and democratic stability. If AI amplifies inequality, economic polarization may intensify. Conversely, if innovation strengthens worker participation, productivity gains could be more broadly shared.
Ultimately, Building Pro-Worker Artificial Intelligence presents a normative and institutional argument: artificial intelligence can either undermine labor or empower it. The outcome depends on deliberate choices in policy design and governance. Rather than treating AI as an autonomous force, the paper calls for shaping technological development in ways that align with economic inclusion and long-term stability.
Reference
Acemoglu, D., Autor, D., & Johnson, S. (2026). Building pro-worker artificial intelligence (NBER Working Paper Series). National Bureau of Economic Research. https://www.nber.org