Henry Farrell on AI as a Social Technology
TL;DR
Political economist Henry Farrell argues that AI systems like large language models function as 'social technologies'—complex institutional mechanisms for processing collective cultural information akin to markets and bureaucracies—rather than as individual agentic intelligences, warning that misunderstanding this distinction creates risks of ideational bubbles when AI narratives collide with reality.
🏛️ AI as Social Technology 2 insights
Institutional information processing systems
Drawing on Herbert Simon, Farrell treats AI not as individual cognition but as complex social information processing comparable to markets, bureaucracies, and democratic institutions that extend human cognitive capacity beyond biological limits.
Compressed cultural knowledge repositories
Large language models function as simplified models of vast quantities of human-produced text, representing frozen cultural information and stereotypical social roles rather than possessing genuine agency or reasoning capabilities.
🎭 The AGI Misconception 2 insights
Massive multiplayer role-playing architecture
Farrell describes LLMs as operating like 'choose your own adventure' games or extended role-play scenarios where users navigate through pre-existing cultural stereotypes rather than interacting with emergent artificial general intelligence.
Rejection of individual agent analogies
This framework directly contrasts with views like Tyler Cowen's that treat AI as comparable to individual human reasoners, arguing that such analogies fundamentally misunderstand the technology's collective, distributional nature.
⚠️ Ideational Bubbles and Historical Context 2 insights
The industrial revolution as the true singularity
Following Cosma Shalizi, Farrell argues that humanity already encountered its 'Lovecraftian' singularity with the Industrial Revolution's creation of vast, inhuman information systems like markets and bureaucracies that we've navigated for centuries.
Risks of narrative collapse
Viewing AI through the lens of social technology reveals the danger of 'ideational bubbles' where hype cycles collapse when AI fails to deliver on AGI promises or when political economies fail to integrate the technology equitably.
Bottom Line
Treat AI as a new institutional mechanism for organizing collective knowledge—similar to how we regulate markets or bureaucracies—rather than as artificial persons, and develop political economic frameworks that account for its role as social infrastructure rather than individual replacement.
More from Conversations with Tyler (Tyler Cowen)
View all
Sajjid Chinoy on Whether India Faces another 1991 Moment
While India is not facing a 1991-style balance of payments crisis, the economy is constrained by weak private investment due to insufficient demand, Chinese import competition, and fiscal pressures from welfare spending crowding out infrastructure investment, necessitating a policy pivot toward employment and exports.
Sajjid Chinoy on Whether India Faces another 1991 Moment
Economist Sajjid Chinoy argues that while India has resolved supply-side constraints and cleaned up corporate balance sheets post-COVID, the economy now faces a binding demand-side crisis exacerbated by the largest energy shock in history, requiring a fundamental policy shift from macro stability toward structural employment generation to trigger private investment.
Emily Chamlee-Wright — 2025 Markets and Society Conference Keynote
Liberal democracy faces collapse from both overt constitutional violations and the 'great forgetting' of its foundational principles. Chamlee-Wright argues that reversing this decay requires repairing liberalism's 'soft tissue'—the shared mental models, automatic norms, and cultural tools that make formal institutions function and scale.
Ritam Chaurey on Placing the Firm at the Center of India’s Structural Transformation
Development economist Ritam Chaurey explains how Indian firms navigate rigid labor regulations by substituting permanent workers with contract labor or exiting formality entirely, revealing that post-1991 liberalization exposed labor laws as the primary binding constraint on manufacturing growth while creating complex, often ambiguous effects on firm productivity and worker welfare.