The Data Economy and Surveillance Capitalism

The internet did not begin as a surveillance apparatus. The early web was largely decentralized, its protocols designed for open information exchange, its business models based on subscription fees, service charges, and display advertising similar to what newspapers sold. Something changed in the early 2000s, and the change was not primarily technological — it was a discovery about what behavioral data could be made to do. That discovery reshaped the internet, built some of the largest companies in human history, and created a set of incentives that now structure how billions of people experience digital life. Understanding the data economy means understanding how that system works, who benefits, and what it produces that its designers did not intend.

How Surveillance Capitalism Works

The scholar Shoshana Zuboff coined the term surveillance capitalism to describe a specific economic logic: the unilateral claiming of human experience as raw material for behavioral prediction products that are sold to business customers. The structure has four components. First, data extraction. A platform provides a free service — search, social networking, maps, email — that gives users a compelling reason to engage. In exchange, the platform captures behavioral data: what users click, how long they look at content, what they search, where they are, who they communicate with, and what they purchase. The service is the bait; the behavioral data is the actual product being harvested. Second, behavioral modeling. The raw behavioral data is processed by machine learning systems to build predictive models. These models answer questions like: what is this person likely to buy in the next thirty days? What content will keep them on the platform longest? What emotional state are they in right now, and what is the optimal ad to show someone in that state? The models are not predicting the future — they are influencing it by shaping the information environment around each user. Third, prediction products. The outputs of the behavioral models are packaged as targeting capabilities and sold to advertisers and other business customers. An advertiser does not buy an ad placement — they buy access to a precisely defined audience: people who have recently searched for competitor products, who are estimated to be in a particular income bracket, and who have shown behavioral signals of a specific emotional state. The precision of the targeting is determined by the quality of the behavioral data and the models trained on it. Fourth, behavioral modification. The platform has an incentive to keep users engaged as long as possible and to produce behavioral states that make users receptive to advertising. This incentive drives product design toward features that maximize engagement: infinite scroll, notification systems designed to create compulsive checking, algorithmic feeds that prioritize content triggering strong emotional reactions. The platform is not neutral — it is actively designing your attention and emotional state because your attention and emotional state are the product it sells.

The Scale and Structure of the Data Economy

The global data economy encompasses several interconnected markets. The largest is digital advertising, where platforms sell targeting capabilities to reach specific audiences. Global digital advertising revenue exceeded $600 billion in 2023, with the majority flowing to Google and Meta, whose surveillance infrastructure is unmatched in depth and reach. The second major market is the data brokerage industry. Companies like Acxiom, Oracle Data Cloud, and hundreds of smaller firms aggregate data from hundreds of sources — loyalty cards, public records, credit transactions, app data sold by developers, data scraped from websites — and sell profiles and audience segments. These firms operate almost entirely outside public awareness. You have never agreed to their terms of service, but they likely have a file on you. The third market is the emerging data infrastructure industry: cloud computing providers that store and process behavioral data, identity resolution services that link your behavior across devices and platforms, and clean room technologies that allow advertisers to match their customer data against platform behavioral data without either party technically seeing the other's dataset. These infrastructure layers make the targeting ecosystem more precise and harder to escape. The combined effect is an economy in which behavioral prediction — knowing what you will do before you do it, and shaping the environment to influence what you do — has become one of the most valuable capabilities on earth.

What Surveillance Capitalism Produces

The data economy has produced genuine goods. Search that is fast and largely accurate. Navigation apps that save hours of driving time. Communication tools that let people maintain relationships across continents. Free access to information that previously required expensive encyclopedias, libraries, and professional consultations. These are real benefits and should not be dismissed. But the same system has produced a set of consistent harms. Algorithmic amplification of misinformation and extremist content, because outrage is engaging. Discrimination enabled by behavioral targeting: research has shown that ad targeting systems have been used to exclude users by race from housing and job ads, that payday loan ads are disproportionately shown to people in financial distress, and that recovery programs can be targeted at people whose behavioral data suggests they are struggling with addiction. The creation of detailed profiles on children who are too young to consent. The sale of sensitive health data — abortion-related searches, mental health app usage, addiction recovery app data — to brokers and insurers. And the fundamental epistemological distortion of an information environment designed not to inform you but to keep you engaged.