Biography — Harvard Business School, the smart machine, the long arc
Shoshana Zuboff was born in 1951. She took her BA in philosophy at the University of Chicago and her PhD in social psychology at Harvard in 1980. She joined the faculty of Harvard Business School in 1981, where she was the Charles Edward Wilson Professor of Business Administration and one of the first tenured women at the school. She retired from active teaching in the early 2000s but remained an active writer; The Age of Surveillance Capitalism (2019) was the culmination of fifteen years of research on the transformations she had been documenting for three decades.
Her first major book, In the Age of the Smart Machine: The Future of Work and Power (1988), was based on extended ethnographic fieldwork in pulp-and-paper mills, telephone companies, and pharmaceutical firms during the first wave of workplace computerization in the 1980s. The central conceptual move of the book — the distinction between automating (substituting machines for human labor) and informating (producing data about work that becomes a new object of management attention) — was prescient enough that the book is still cited by labor scholars studying contemporary AI deployment in the workplace.
Between 1988 and 2019 Zuboff published two more books — The Support Economy (2002, with James Maxmin) and several edited volumes — butThe Age of Surveillance Capitalism is the work the framework borrows from. The book took a decade to write, drew on extensive documentary research into Google's and Facebook's patent filings and financial disclosures, and arrived in 2019 as the most comprehensive single-author analysis of the business model that had quietly become dominant during the decade preceding it. Zuboff is now a professor emerita at HBS and continues to publish and speak on the surveillance-capitalism framework and its successors.
The conceptual machinery
The Age of Surveillance Capitalism is built around a small number of carefully defined terms that Zuboff coined for the book and that have since entered general use. The framework borrows four of them as load-bearing.
The data about human activity that exceeds what is needed to provide the surface service. Google's search service can be provided with the queries; the location data, click patterns, scroll velocity, dwell times, and all the other behavioral exhaust that the service collects beyond the queries themselves is the surplus. Zuboff's argument is that the realization (around 2001–2004 at Google) that this surplus could be extracted and monetized was the founding move of the surveillance-capitalism business model.
Behavioral surplus is fed into machine-learning systems that produce prediction products — forecasts of future behavior that are then sold to actors whose interest is in shaping that behavior. Advertisers were the first market for these products, but the framework has since extended to insurance, credit, hiring, real estate, and political persuasion. The crucial structural point is that the prediction products are sold to parties other than the people whose behavior is being predicted, and often without the knowledge or consent of those people.
The mechanism by which behavioral surplus is extracted is unilateral. Users do not meaningfully consent to the extraction in any way that ordinary contract law would recognize as consent — the terms-of-service agreements that purport to authorize the extraction are not read, are not negotiable, and would not be defensible on most informed-consent standards. Zuboff's term for this isthe unilateral incursion into experience: the surveillance-capitalism business model is an unauthorized seizure of human experience as raw material, comparable in structure (though not in degree) to the historical primitive accumulations that produced earlier capitalist orders.
Zuboff distinguishes instrumentarian power — the power to predict and shape behavior at scale — from oldertotalitarian power, which sought to dominate through ideology and coercion. Instrumentarian power does not require ideological commitment; it requires only behavioral compliance, achieved by making certain behaviors easier and others harder through interface design. The framework borrows the distinction because it names a specific mechanism that the older vocabulary of authoritarianism cannot capture.
Temporal influences — the long-arc moment of the book
The decade of research
The book's timing is not incidental. The decade Zuboff spent researching it (roughly 2009–2019) was the decade in which surveillance capitalism consolidated as the dominant business model of the digital economy. Google's transition from search-as-product to advertising-as-product had happened earlier (2001–2004); Facebook's emergence as the second great surveillance-capitalism firm occurred in the late 2000s; the smartphone (after 2007) extended the extraction surface from the desktop browser into continuous mobile experience; the mid-2010s consolidation of recommender-system-driven feeds (Facebook, Instagram, YouTube, TikTok) produced the attention-economy critique that became mainstream by the late 2010s. The book is the synthesis of this decade, by a researcher who had been watching it carefully throughout.
Disciplinary & institutional position
Zuboff worked from inside Harvard Business School — institutionally the heart of mainstream American management research. The book's critique of contemporary tech-industry business models is therefore unusual in its source: not from the academic Left, not from civil society, not from competitor firms, but from an HBS chair. This position has costs and benefits. The cost is a certain stylistic conservatism (the book is long, formal, and methodologically careful in ways some readers find difficult). The benefit is institutional credibility — the argument is much harder to dismiss as ideological when it is being made by someone with Zuboff's credentials.
Intellectual lineage
The book sits in conversation with several traditions: the long arc of critical theory on instrumental reason (Adorno, Horkheimer); Polanyi's The Great Transformation on the political construction of market society; Hannah Arendt on totalitarianism (with the distinction Zuboff draws specifically against); and the technical literature on behavioral economics, recommender systems, and machine learning. Zuboff's ability to read the technical primary sources (patent filings, financial disclosures, Google internal documents obtained through litigation) is what gives the book its evidentiary base; her ability to integrate this reading with the political-theory tradition is what gives it its analytic depth.
Reception and the AI inflection
The book's 2019 release was followed by substantial public engagement — major reviews in the Anglophone press, congressional citations, broad uptake in tech-policy circles. The post-2022 AI moment has, somewhat surprisingly, both extended Zuboff's argument and partly displaced it. Extended, because the underlying mechanism (extraction of behavioral surplus, sale of prediction products to third parties) applies straightforwardly to generative AI training data. Displaced, because the doom-and-hype discourse around AI has largely overrun the more measured surveillance-capitalism critique in public attention. The framework should note both — that Zuboff's analysis is still load-bearing for understanding the substrate, and that the discourse displacement we critique elsewhere is partly visible as the displacement of her argument.
What LeResearch specifically borrows
Operational concepts:
- ·Behavioral surplus and prediction products as the mechanism
§5's recommender-systems bullet borrows this whole vocabulary. Hiring algorithms, lending models, content recommendation systems, dating apps, news feeds — all run on the same underlying mechanism Zuboff named for advertising. The framework owes Zuboff the precise technical vocabulary for what these systems are doing.
- ·The shallow-gradient-as-strategy reading
The framework's §3 normalization-gradient argument is reinforced by Zuboff's careful demonstration that the shallow gradient was not accidental. The unilateral incursion into experience was deliberately structured to avoid triggering the sensors that would have produced public resistance. This is the most important specific case the framework has of the slow-gradient pattern as designed strategy, not just emergent dynamic.
- ·Instrumentarian power as named pathology
The distinction between instrumentarian and totalitarian power gives the framework vocabulary for the specific contemporary political mechanism that does not require ideology, only behavioral compliance achieved through interface design. This is directly relevant to §6 (compression and silent versioning) and to the broader question of what AI mediation does to democratic deliberation.
- ·The corporate-choice analytic posture
Zuboff insists that the present arrangement is the consequence of specific corporate choices made by specific people for specific reasons, and that those choices are reversible. Resist both the determinist reading (technology was going to do this) and the libertarian reading (users freely chose). Hold the corporate choosers accountable in the way one holds any actor accountable for the consequences of their choices. The framework borrows this entire posture.
Background posture: the documentary discipline. The book's evidentiary base is patent filings, financial disclosures, internal Google documents, and careful readings of corporate communications against their own subsequent actions. The framework owes Zuboff the recognition that this documentary register is what makes the critique authoritative, and that arguments at similar scope without similar evidence should be more conditional.
What we set aside
- ·The book's length and stylistic register
The Age of Surveillance Capitalism is over 700 pages and stylistically formal in ways that limit its readership. The framework borrows the analytic vocabulary while writing in a tighter register that is more compatible with the publishing economy LeResearch actually operates in.
- ·The strong primitive-accumulation analogy
Some of the book's framing draws extended analogies between contemporary behavioral surplus extraction and the historical primitive accumulations that produced capitalism (enclosures, colonization). The framework finds the analogy productive in outline but is wary of pushing it too far — the differences in degree and mechanism matter, and the analogy can crowd out specific contemporary analysis.
- ·The somewhat US-centric empirical focus
The book is overwhelmingly about American tech firms, American regulatory failures, and American cases. The framework wants the vocabulary to travel to other contexts (European regulation, Chinese state-corporate fusions, Indian platform economies) where Zuboff's specific cases do not apply directly.
- ·The prescriptive register
The book's policy prescriptions tend toward strong national-regulatory responses (new data-rights frameworks, antitrust enforcement, the GDPR taken further). The framework borrows the diagnostic without committing to a specific regulatory programme. There are several political horizons compatible with the underlying analysis.
What we still owe — the deeper unresolved
Three open questions, in increasing order of importance.
Does generative AI extend or transform the surveillance-capitalism mechanism?
Zuboff's central case was the second decade of the twenty-first century — search and social as surveillance economies, behavioral surplus extracted from web activity. Generative AI extends the extraction surface to the conversation itself: every question asked, every draft revised, every document edited becomes available behavioral data of a kind earlier surveillance regimes could not have produced. Whether this is a continuation of the Zuboff argument or a phase change that requires new vocabulary is the open question the framework is most actively working through. The first read is straightforward extension; the proper read requires care we have not yet given.
What replaces the unilateral incursion when the user is voluntarily prompting?
The unilateral-incursion framing assumes a user who has not asked for the extraction. LLM-mediated work is a different case — users are actively volunteering input, often sensitive or proprietary input, in exchange for assistance. The relationship is not consensual in the strong sense (terms of service still apply, the data still goes into training sets), but it is also not obviously the same as the background extraction Zuboff documented. The framework owes a treatment of what the consent-and-extraction relationship actually looks like in this case, and of what the political stakes of the difference are.
The instrumentarian-power question for the AI moment
Zuboff's deepest argument — that instrumentarian power is a specifically twenty-first-century pathology that requires new political vocabulary — is what the framework most needs to extend. AI deployment in institutions is producing arrangements in which the institution does not require ideological commitment from its members, only behavioral compliance achieved through interface design and algorithmic decision-support. What instrumentarian power looks like inside courts, schools, hospitals, and government agencies as they absorb AI tools is the question §5 and §6 together are reaching for, and it is the deepest open question this whole open-threads series has produced. The Zuboff vocabulary is necessary; we are still working out what it specifically allows us to say.
Where to start, if you are reading her for the first time
- ·The Age of Surveillance Capitalism (2019)
The major work. 700+ pages but well organized; start with Part I (the discovery of behavioral surplus and the founding of the business model at Google) and Part III ( instrumentarian power and the political implications). Parts II and IV can be skimmed or returned to selectively.
- ·In the Age of the Smart Machine (1988)
The early book on workplace computerization. The automate vs. informate distinction is foundational and applies directly to contemporary AI deployment in workplaces. Useful for seeing the long arc of Zuboff's thinking on technology and labor.
- ·Surveillance Capitalism or Democracy? (2022 article in New Labor Forum)
A short, accessible synthesis of the argument. Useful as an entry point if the full book is too long for the present purpose.
- ·Zuboff in interviews and lectures (2019 onward)
Multiple long-form interviews with Zuboff are available on the Berggruen Institute, On The Media, and similar venues. Useful for understanding which parts of the framework she emphasizes when speaking to general audiences.
- ·The companion threads on Graeber
Graeber on Bullshit Jobs and Graeber on Debt — together with Zuboff, these form the labor-side substrate beneath §5. The three books in conversation with each other are the framework's core reading list for the contemporary labor-and-AI question.