You On AI Field Guide · Shoshana Zuboff The You On AI Field Guide Home
TxtLowMedHigh
PERSON

Shoshana Zuboff

The scholar who documented the silent transformation of every computerized workplace—and who named the economic system that turned human experience itself into raw material for machine intelligence.
Shoshana Zuboff is the cartographer of the modern worker's invisible wound. In the early 1980s, she embedded herself inside paper mills and insurance companies being computerized, watching as the knowledge workers had built in their hands and bodies was displaced by screens—a transition she captured in the foundational distinction between automating and informating. Automation eliminates human labor; informating generates new data about processes that, if distributed equitably, could create new forms of understanding. The tragedy Zuboff documented across four decades is that organizations consistently chose automation without informating—taking the cost savings and leaving the informating dividend unrealized. By 2019, she had named a deeper pathology: surveillance capitalism, the economic system in which human experience is claimed as free raw material, fed into AI factories, and sold as prediction products to businesses that want to modify behavior rather than serve it. Her framework illuminates the AI moment with uncomfortable precision: the builders celebrating expanded capability in [YOU] on AI are simultaneously generating behavioral surplus—cognitive signatures, creative rhythms, professional judgment patterns—that the platforms extract and monetize without the builder's awareness or consent.
Shoshana Zuboff
Shoshana Zuboff

In the [YOU] on AI Field Guide

The Field Guide asks what it truly costs to take the orange pill—to accept expanded capability, compressed timelines, the twenty-fold productivity multiplier. Zuboff supplies the most rigorous available accounting of what the metrics exclude. When Edo Segal describes watching engineers in Trivandrum accomplish in days what once required weeks, Zuboff's framework asks: who is sustaining those engineers during the sprint? What household labor absorbs their extended hours? What embodied knowledge is being displaced alongside the implementation friction? And what cognitive behavioral data is flowing, invisibly, from every prompt session into the platform's training apparatus?

Her automating-versus-informating distinction reshapes the book's central question. Segal argues that AI frees workers from implementation drudgery and elevates them to judgment, vision, and creative direction—the ascending friction thesis. Zuboff's empirical record in the paper mills tells the same story with a darker second act: the informating potential was real, but the institutional investment required to realize it was rarely made. Organizations captured the cost savings and left workers with thinner skills, weaker authority, and no compensation for the embodied knowledge they had surrendered. The AI moment is either the exception to this pattern or its most spectacular confirmation, and which it becomes is a question not of technology but of institutional design.

Automating vs. Informating
Automating vs. Informating

Her concept of evaluative intellective skill supplies the sharpest available diagnosis of the skill gap the AI transition creates. The paper mill worker who moved from the floor to the control room needed new cognitive capacities to interpret digital representations of the process she had previously touched. The knowledge worker who moves from writing code to evaluating AI-generated code needs an analogous shift—but the evaluative capacity depends on the constructive foundation that the tool's frictionless efficiency is simultaneously eroding. You need to have built to evaluate well what the machine builds, and the machine is eliminating the building.

The surveillance capitalism framework completes the Field Guide's most uncomfortable chapter. The builder who works with Claude at three in the morning, ideas connecting through the warm fluency of the conversation, is experiencing genuine collaboration and generating behavioral surplus simultaneously. Every prompt reveals cognitive architecture. Every revision reveals values. Every abandoned direction reveals judgment. The extraction is structural, not conspiratorial, and the builder who is honest about the collaboration's gifts must also be honest about what the collaboration takes.

Ascending Friction Thesis
Ascending Friction Thesis

Origin

Born in 1951 and educated at the University of Chicago and Harvard, where she eventually became the first woman to hold an endowed chair at Harvard Business School, Zuboff began her career as a scholar of work and consciousness. Her doctoral research asked a question that seemed purely philosophical but turned out to be urgently empirical: what happens to the human mind when the relationship between worker and work is fundamentally transformed by technology? To answer it, she spent years inside organizations that were being computerized in the early 1980s—paper mills in the American South, insurance companies in New England, telecommunications firms—conducting the kind of sustained ethnographic fieldwork that most management scholars had abandoned.

Surveillance Capitalism
Surveillance Capitalism

What she observed was the destruction of action-centered skill: knowledge built through physical engagement with material processes, living in nerve endings calibrated over decades, irreplaceable and irreproducible through symbolic means. The paper mill worker who could feel the consistency of pulp between his fingers possessed expertise that no screen could replicate. When the screen replaced the feel, the expertise had no substrate in which to persist. Zuboff documented the workers' response: not resistance to technology but grief for a way of knowing. Her 1988 masterwork In the Age of the Smart Machine named the dynamic and asked whether the informating potential of computerization—the genuine new knowledge that digital systems could create—would be distributed to empower workers or captured to eliminate them.

Behavioral Surplus
Behavioral Surplus

The answer, observed across four decades and codified in the 2019 Age of Surveillance Capitalism, was more troubling than she had feared. The platforms that dominated the digital economy had discovered something the paper mills had not imagined: that the most valuable product of computerization was not accelerated production but behavioral data. The surveillance apparatus was not incidental to the business model. It was the business model, and artificial intelligence was its engine.

The Informating Dividend
The Informating Dividend

Key Ideas

Automating versus informating. Every technology that interposes a layer of abstraction between worker and work simultaneously automates (eliminating some human labor) and informates (generating new data about the process). The automation is captured immediately as productivity gain. The informating potential—the new knowledge that could elevate rather than merely displace workers—requires institutional investment to realize. Zuboff's four-decade finding is that the investment is rarely made. The cheaper path prevails: automation without informating, displacement without elevation. The informating dividend is real. Whether it flows to workers depends on institutional choice, not technological necessity.

The Worker's Dilemma
The Worker's Dilemma

The three cognitive transitions. Zuboff maps the history of human-machine relations as three qualitative shifts. First, from touching to reading: the paper mill worker who operated the digester by hand migrates to a control room, exchanging embodied knowledge for intellective skill—the capacity to construct mental models from symbolic representations. Second, the AI era introduces a shift from reading to conversing: the worker no longer interprets raw data but evaluates interpretations the machine has already made. The cognitive demand shifts from making sense to judging sense already made. The third transition demands evaluative intellective skill whose foundation—constructive practice—the tool is simultaneously eroding.

Evaluative Intellective Skill
Evaluative Intellective Skill

Behavioral surplus and the surveillance factory. Behavioral surplus is the data that exceeds what is needed to provide the service the user requested—the patterns, rhythms, and signatures of human experience claimed by platforms as free raw material. Zuboff traces its path: scraped from human activity, processed in AI's computational factory, sold as prediction products to parties whose interest is modification, not service. AI interactions produce qualitatively richer behavioral surplus than any previous digital interaction, because they reveal cognitive architecture itself—how a person thinks, not merely what they search for.

The panoptic sort and epistemic inequality. The panoptic sort classifies populations by categories that determine opportunity. In the AI era, the sorting criterion has become cognitive—who can evaluate machine output with sufficient judgment, who can formulate problems worth solving, who possesses the strategic vision to direct amplification toward genuine value. This new epistemic hierarchy compounds: better evaluation produces better outcomes, which builds more domain knowledge, which improves evaluation further. The gap widens with each iteration, and the mechanism is invisible to those being sorted.

Institutional design as the only answer. Zuboff refuses fatalism. The informating dividend can be realized, but only through institutional structures designed to realize it: enforceable standards for practice preservation, governance frameworks for the extraction of behavioral surplus, transparency requirements for the sorting mechanisms, and public investment in the evaluative intellective skill that the AI transition demands but does not develop. Individual organizational choices—one leader's decision to keep the team, one researcher's structured pause framework—are necessary and insufficient. The forces operate at the level of the market. The counter-movement must operate there too.

Debates & Critiques

The central debate is whether the informating dividend will be realized at scale before the automation displacement renders the question moot. Segal argues for depth—that the ascending friction thesis describes a genuine elevation, that judgment and architectural instinct survive and indeed flourish when implementation friction is removed. Zuboff's empirical record introduces the institutional complication: the pattern across every previous smart machine transition is that the potential was real and the realization was rare. Byung-Chul Han's critique of smooth culture runs parallel to Zuboff's without sharing her institutional optimism—Han sees the trajectory as one of increasing subjugation to the achievement imperative, with no institutional counter-movement capable of redirecting it. A second disagreement concerns the surveillance capitalism framework itself: critics including Cory Doctorow argue that behavioral modification—the ability of prediction products to actually shape what people do—is substantially “snake oil,” that the real problem is monopoly power rather than epistemic manipulation, and that antitrust enforcement rather than abolition is the adequate response. Zuboff counters that AI-era behavioral surplus is qualitatively different from its social media predecessors, because cognitive behavioral data reveals the architecture of thought itself rather than merely its surface expressions. The deepest open question her work leaves is whether democratic institutions can build the counter-movement fast enough—Zuboff herself noted in 2024 that the percentage of the world's population living in democracies fell from 51 to 28 percent between 2004 and 2024, and that she sees causality in that decline.

The Three Cognitive Transitions

Zuboff's map of how AI remade the worker's relationship to knowledge
First Transition · Touching
Action-Centered Skill
Knowledge that lives in the body's calibrated engagement with material. The paper mill worker who could feel the pulp. The surgeon who could tell by the resistance of tissue. Skill that cannot survive the removal of the body from the process—and that no screen can replace.
Second Transition · Reading
Intellective Skill
The capacity to construct mental models from symbolic representations. The control room worker who interprets digital displays of the process she once touched. More precise in some dimensions; permanently severed from the embodied knowledge that made critical evaluation possible.
Third Transition · Conversing
Evaluative Intellective Skill
Judging the quality of sense the machine has already made. The hardest of the three demands, because the machine's output is optimized for plausibility rather than accuracy—and detecting where the two diverge requires precisely the constructive experience the tool is eliminating.

Further Reading

  1. Shoshana Zuboff, In the Age of the Smart Machine: The Future of Work and Power (Basic Books, 1988)
  2. Shoshana Zuboff, The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power (PublicAffairs, 2019)
  3. Shoshana Zuboff, “The Coup We Are Not Talking About,” The New York Times (2021)
  4. Shoshana Zuboff, interview in El País, “AI is surveillance capitalism continuing to evolve and expand” (December 2025)
  5. Xingqi Maggie Ye & Aruna Ranganathan, “How AI Changes Knowledge Work,” Harvard Business Review (February 2026)
Explore more
Browse the full You On AI Field Guide — over 8,500 entries
← Home0%
PERSONBook →