Generative AI could erode the bargaining power of elite cognitive workers, opening political space for institutional redistribution; a revival of institutional-economics thinking helps explain when technological progress might enable progressive change.

Main Finding

The diffusion of generative AI (Gen-AI) threatens to displace not routine jobs but many nonroutine cognitive, highly educated occupations that previously held strong bargaining power. This erosion of scarcity and exclusivity among influential workers could broaden support for institutional reforms that redistribute gains from rising output as human labor input falls, reviving insights from Original Institutional Economics about how technological change can catalyze progressive institutional shifts by altering the balance between property and economic power.

Key Points

• Historical context: Prior technological waves were largely routine- or skill-biased, causing job polarization and concentrating labor-market power among highly paid, nonroutine cognitive workers.
• Gen-AI differs qualitatively: it is most disruptive in the very occupations (nonroutine cognitive roles) that historically enjoyed growing bargaining power.
• Political economy implication: When technological displacement threatens influential segments of the labor force, political coalitions and societal support for redistribution and institutional reform become more likely.
• Theoretical framing: The paper applies Original Institutional Economics to distinguish property rights from economic power and to identify conditions under which technology can enable institutional change.
• Normative suggestion: Gen-AI’s diffusion could open a window for progressive reforms (e.g., stronger redistribution, labor protections, governance measures) because the politically powerful group may lose its ability or incentive to block such changes.

Data & Methods

• Nature of analysis: Conceptual and theoretical; the paper synthesizes existing empirical patterns from prior technological waves and frames a new argument about Gen-AI’s likely social and political effects.
• Evidence base: Comparative-historical reasoning about past labor-market responses to technology (routine- and skill-biased change), plus application of institutionalist theory (Original Institutional Economics) to interpret political-economic dynamics.
• Empirical contribution: No new microdata or econometric tests are reported in the abstract; the contribution is analytic and hypothesis-generating.
• Suggested empirical approaches (implicit): To test the arguments, one could analyze occupation-level automation exposure, wage and employment trajectories in nonroutine cognitive jobs, shifts in political preferences and coalition formation among affected workers, and changes in institutional outcomes (tax, labor law, social insurance) across jurisdictions with differing Gen-AI adoption.

Implications for AI Economics

• Redistribution and policy windows: AI economists should anticipate and model not only labor-market displacements but also endogenous institutional responses—tax policy, social insurance expansion, labor regulations—that may follow when high-status workers are affected.
• Rethinking power dynamics: Analyses of AI’s economic effects must consider changing bargaining power across occupations, not just aggregate productivity or employment elasticities.
• Broader equilibrium effects: Institutional reforms prompted by Gen-AI diffusion could alter capital–labor splits, incentives for further automation, and diffusion patterns—feedbacks that should be incorporated into long-run AI-economy models.
• Research priorities: Empirical work on who is displaced by Gen-AI (occupation/skill level), measurement of political and institutional responses, and formal models linking technological diffusion to shifts in economic power and policy outcomes.
• Policy design: Proactive design of redistributive and governance mechanisms (progressive taxation, portable benefits, worker voice/representation, AI-specific regulation) may be both necessary and politically feasible as Gen-AI affects influential worker groups.