Power and Progress by Daron Acemoglu and Simon Johnson: Study & Analysis Guide

Technological change is often hailed as an inevitable engine of human betterment. In Power and Progress, Daron Acemoglu and Simon Johnson present a profound counter-narrative: the direction of innovation and who benefits from it are not predetermined by engineering possibilities but are instead a function of social and political power. Their sweeping thousand-year argument culminates in a stark warning about artificial intelligence. Understanding their framework is crucial for anyone grappling with why, in an age of stunning technological prowess, broad-based prosperity feels increasingly fragile.

The Central Thesis: Progress is a Political Choice

The book’s foundational argument is that technological progress does not automatically produce shared prosperity. Acemoglu and Johnson reject the notion of a "productivity bandwagon," where gains from new machines naturally trickle down to all of society. Instead, they insist that the direction of innovation—what problems technology seeks to solve and whose tasks it aims to augment or replace—is shaped by the prevailing balance of power. When elites control the agenda, innovation tends to be labor-replacing and sustains or widens inequality. Broadly shared benefits only occur when countervailing institutions—like unions, democratic presses, and responsive governments—force a redirection of innovation toward more productivity-enhancing paths that create new tasks and opportunities for the majority. This framework shifts the debate from "how much" technology to "what kind" and "for whom."

A Millennium of Contested Technology

To ground their thesis, the authors conduct a historical tour de force, demonstrating how elite capture of technology is the rule, not the exception. Medieval advances in the plow and three-field system primarily enriched feudal lords, not the peasants whose labor was essential. The critical case study is the British Industrial Revolution. Early textile machinery, like the spinning jenny and power loom, was explicitly designed to deskill and control workers, not to make them more productive. The result was decades of "immiserizing growth"—rising national output paired with falling wages and brutal working conditions for the masses.

The shift toward broadly shared gains, they argue, began only in the later 19th and 20th centuries. This was not a gift from benevolent industrialists but was forged through intense conflict. The rise of countervailing power—including organized labor, investigative journalism, progressive taxation, and eventually the modern regulatory state—forced a reorientation. Innovation began to focus more on creating new, higher-paying jobs (e.g., in engineering, management, and services) and improving general welfare through public health and education. This period demonstrates their core lesson: shared progress requires a continual struggle to reshape the institutional environment that guides technological change.

The Modern Parallels: From Automation to AI

Acemoglu and Johnson see troubling echoes of the early Industrial Revolution in recent decades. The dominant trajectory of innovation since the 1980s, they contend, has shifted back toward automation and labor-replacing technologies, driven by a corporate philosophy that prioritizes cost-cutting and shareholder value over job creation. This has contributed significantly to wage stagnation, inequality, and the decline of the labor share of income. Digital technologies, while transformative, have often been deployed to surveil workers, automate routine tasks, and create "fissured" workplaces without a corresponding boom in new, good jobs.

This analysis sets the stage for their urgent examination of artificial intelligence. They warn that AI is currently on a path of "wrong direction" innovation. Its development is heavily influenced by a concentrated tech elite and is focused on areas like facial recognition, algorithmic management, and predictive policing, which can exacerbate inequality and surveillance. Without a deliberate shift, AI risks becoming the ultimate tool for elite control and worker displacement, cementing a new "digital feudalism" rather than ushering in a new age of shared prosperity.

Redirecting Technology: The Institutional Imperative

The book’s prescription moves beyond simple redistribution. The primary goal must be to redirect innovation itself toward more human-complementary technologies. This means designing AI that augments human skills and creates new tasks and industries, rather than simply replacing checkout clerks, drivers, or diagnostic radiologists. Achieving this requires rebuilding countervailing institutions capable of steering the tech agenda.

They propose several levers: revitalizing antitrust policy to break up concentrated tech power, reforming corporate governance to consider worker and community interests, and using public funding and procurement to prioritize human-complementary AI research. Crucially, they argue for a more vigorous and confident democratic state to set the rules of the game, akin to its role in shaping the inclusive direction of mid-20th century innovation.

Critical Perspectives on the Argument

While the historical sweep and institutional analysis are characteristically rigorous and persuasive, the policy prescriptions for redirecting AI innovation remain somewhat abstract. Critics might argue that defining and governing "human-complementary" AI is extraordinarily complex in practice. The book also presents a largely Western narrative of technological contestation; a deeper global perspective could enrich the analysis. Furthermore, some may find the portrayal of the tech elite as a monolithic, extractive force overlooks internal debates and genuine belief in the beneficial potential of AI among some developers. The central challenge the book leaves us with is operational: how to build the necessary political coalitions and institutional capacity to enact their vision in a polarized and globalized world.

Common Pitfalls

1. Misreading the Thesis as "Anti-Technology": A common mistake is to interpret the book as a Luddite tract. It is not opposed to technology but is fiercely critical of who controls its direction. The argument is for smarter, more democratic steering, not for halting progress.
2. Overlooking the "Direction" in Favor of "Pace": Discussions of technology often focus on speed (e.g., "exponential change"). This book demands we focus first on vector—the social and economic goals toward which innovation is aimed.
3. Equating Countervailing Power Only with Unions: While unions are a classic example, the authors' concept of countervailing institutions is broader. It includes a free press, an independent judiciary, effective regulatory agencies, and vibrant civic organizations—any institution that can check concentrated elite power.
4. Dismissing the AI Warning as Alarmist: It is easy to see the current path of AI as inevitable. The book's historical service is to show that today's tech trajectory is a choice, and past choices with similar power dynamics led to generations of suffering before being corrected.

Summary

• Technology is not autonomous: The path of innovation is shaped by politics, power, and pre-existing institutions, not just scientific possibility.
• The default is elite capture: History shows that without a balancing force, elites will steer technology to increase their wealth and control, often at the expense of broader welfare—a pattern seen from feudal agriculture to early factory automation.
• Shared prosperity requires struggle: The broadly beneficial "productivity bandwagon" of the mid-20th century was a hard-won anomaly, achieved through the rise of unions, democracy, and regulation that forced innovation onto a more inclusive path.
• AI is at a crossroads: Current AI development, dominated by a narrow set of corporate and intellectual interests, risks repeating the extractive patterns of the past, leading to greater inequality and surveillance.
• Redirecting innovation is key: The solution is not just to redistribute the gains from AI but to fundamentally redirect its development toward human-complementary technologies that create new opportunities and tasks for workers.
• Institutions must be rebuilt: Achieving this requires revitalizing countervailing power—through antitrust, corporate governance reform, and public investment—to democratize the control over humanity's technological future.