Daron Acemoglu

In the AI Story

Acemoglu's collaboration with Allen on 'How AI Fails Us' brought together economics and political theory in an analysis of the structural antidemocratic tendencies of contemporary AI development. The paper's core argument—that the dominant AI paradigm 'tends to concentrate power, resources, and decision-making in an engineering elite'—combines Acemoglu's institutional analysis of technology with Allen's theory of democratic equality.

Acemoglu's 2024 Nobel Prize in Economics, shared with Robinson and Johnson, recognized the institutional analysis of long-run economic development. The award's timing—in the middle of the AI boom—signaled the relevance of the institutional framework for understanding contemporary technology development. Acemoglu has been outspoken about the distributional risks of AI, arguing that the technology's benefits will concentrate narrowly unless deliberate institutional choices direct them broadly.

Power and Progress (2023), co-authored with Simon Johnson, extends the institutional analysis to the history of technology from the medieval period to the present. The book argues that the narrative of technology as an unambiguous source of broadly shared prosperity is historically false: most technological transitions have concentrated benefits narrowly in the short and medium term, with broad sharing emerging only when specific institutional conditions—worker organization, democratic governance, redistributive mechanisms—were in place.

Acemoglu's work provides the economic foundation for Allen's argument that the AI transition requires specific institutional innovations to distribute its benefits broadly. The institutional vacuum that destroyed the Luddites threatens to destroy the workers displaced by AI, absent deliberate institutional choices to protect them. Acemoglu has been explicit that this is not an argument against AI but an argument for democratic governance of AI's deployment.

Origin

Daron Acemoglu is Institute Professor at MIT and has been among the most influential contemporary economists working on the political economy of development and technology.

Key Ideas

Technology choices are institutional choices. The distributional consequences of technology are shaped by institutional arrangements, not determined by the technology itself.

Concentration as default. Historically, technological transitions have concentrated benefits narrowly in the short and medium term.

Broad sharing requires institutions. Technologies produce broadly shared prosperity only when specific institutional conditions are in place.

AI institutional vacuum. The current AI transition is occurring in an institutional vacuum that threatens to reproduce the Luddite pattern.

Democratic governance imperative. The alternative to AI's concentration of benefits is deliberate democratic governance of its deployment.