Surveillance capitalism is an economic system built on the extraction of personal data from everyday digital activity, which is then processed and sold to predict and influence human behavior. The term, popularized by scholar Shoshana Zuboff, describes a logic that has come to dominate much of the internet economy: companies offer free or cheap digital services, harvest far more data than they need to improve those services, and turn that surplus into products sold to advertisers and other business clients who want to reach, persuade, or nudge specific people. What makes this model distinct from older forms of corporate data use is both its scale and its ambition. It does not just want to know what you have done; it wants to know what you will do next, and increasingly, to shape that future action.
How the Extraction Works
Every tap, scroll, pause, and search query you produce while using a phone or computer generates data. Some of that data genuinely helps improve the product you are using. If a mapping app learns that a particular route is congested, it needs real-time location data to route you around traffic. If a music service notices you skip certain songs, it needs listening data to recommend better ones. That functional slice of your data is not what critics worry about.
The concern starts with everything collected beyond what is needed for the service. Your location tracked continuously even when the map is closed. The accelerometer in your phone recording how quickly you walk. The pattern of which apps you open before and after checking your bank balance. The ambient audio a smart speaker picks up while it waits for a wake word. This excess is what Zuboff calls “behavioral surplus,” and it is the raw material of surveillance capitalism. It gets fed into machine-learning systems that build predictive models of individual behavior, which are then packaged and sold, primarily to advertisers, but also to insurers, employers, political campaigns, and data brokers who resell the profiles downstream.
The business logic is straightforward: the more data you collect, the more granular your behavioral predictions become, and the more money those predictions fetch. This creates a relentless incentive to gather data from more places, more often, in more intimate detail. It also means the companies best positioned to extract this data, the ones whose products sit at the center of daily life, become enormously powerful. Search engines, social networks, and mobile operating systems are not incidental winners in this model. They are structurally dominant because of how much behavioral surplus flows through them.
Targeted Advertising as the Revenue Engine
Advertising is the primary market where behavioral predictions are monetized. The shift from traditional advertising to data-driven targeting has been dramatic. In the old model, an advertiser placed an ad in a newspaper or during a television show and hoped the right people saw it. In the surveillance-capitalist model, the advertiser specifies exactly who should see the ad, sometimes down to individual-level characteristics like recent purchases, emotional state inferred from social media activity, or physical location. Research into mobile advertising has explored unified modeling frameworks that combine machine-learning targeting with auction models to examine how market outcomes change under different targeting regimes, underscoring how precise ad delivery has become a technical and economic science in its own right.1Marketing Science. Targeting and Privacy in Mobile Advertising
This precision is what makes the model so lucrative. An ad shown to someone already shopping for running shoes converts at far higher rates than a billboard on a highway. But it is also what makes the model so invasive: the precision depends on knowing enough about the person to predict what they want, sometimes before they know it themselves. The data pipelines that feed this system touch almost every corner of digital life, from retail loyalty cards and fitness trackers to browser cookies and email metadata.
The architecture of targeted advertising has prompted a parallel field of research into privacy-preserving alternatives. Researchers have catalogued and compared various techniques designed to let ad targeting function without exposing individual-level data, exploring different underlying architectures, privacy mechanisms, and deployment scenarios.2arXiv. Privacy in Targeted Advertising: A Survey These range from on-device processing that never sends raw data to a server, to differential-privacy methods that add statistical noise so individual records cannot be reconstructed. None of these approaches has displaced the dominant model yet, but the technical groundwork exists for advertising systems that do not require a panopticon to function.
Digital Nudging and the Modification of Behavior
Prediction is only half the business. The other half is influence. Once a system knows your behavioral patterns well enough, it can intervene at precisely the right moment to steer your choices. This goes beyond showing you a relevant ad. It includes the design of notification timing, the sequencing of content in your feed, the placement of buttons and default settings, and dozens of other interface decisions that subtly guide what you do next.
The technique is called digital nudging, and it is not inherently sinister. Research on e-learning platforms, for instance, has found that digital nudging can enrich learning experiences by positively influencing how working professionals interact with educational tools.3Education and Information Technologies. Examining the impact of digital nudging on user interaction with E-platforms: a study of working professionals in distance learning A well-placed reminder to complete a course module or a progress bar that makes your effort visible can be genuinely helpful.
The trouble begins when the same techniques are deployed not to help you accomplish your own goals, but to serve someone else’s. A social media platform nudging you to spend one more minute scrolling, a shopping app surfacing a limited-time offer when your browsing data suggests you are on the fence, a political campaign micro-targeting you with content designed to trigger a specific emotional response: these are all applications of the same behavioral modification toolkit. The line between a helpful nudge and a manipulative one is often invisible to the person being nudged, which is part of what makes the practice so contested.
When Surveillance Follows You to Work
Surveillance capitalism is not confined to consumer technology. Its logic has migrated aggressively into the workplace. Employers have always monitored workers to some degree, from factory floor supervisors counting output to call centers timing phone conversations. But the current generation of workplace surveillance tools goes far beyond productivity tracking.
Research on the connected workplace has documented how modern employee monitoring extends the scope of data collection well past what is needed for performance appraisals or managerial oversight. This includes treating an employee’s body as a data source, disrespecting the boundaries between work and private life, and using the gathered information for subtle persuasion, manipulation, and coercion.4PubMed Central. The connected workplace: Characteristics and social consequences of work surveillance in the age of datification, sensorization, and artificial intelligence Wearable devices that track movement and biometrics, keystroke-logging software, webcam-based attention monitoring during remote work, and even analysis of tone and sentiment in internal communications have all become part of the employer’s toolkit.
Algorithmic management systems, which automate scheduling, task assignment, and performance evaluation, deepen this dynamic. A synthesis of research on algorithmic management identifies it as functioning simultaneously as a surveillance and control system, raising concerns about pervasive monitoring, bias, dehumanization, and worker alienation, even as it enhances coordination and efficiency.5New Technology, Work and Employment. The Rise of Algorithmic Management and Implications for Work and Organisations Gig-economy drivers, warehouse workers, and remote freelancers all experience versions of this: their work is quantified, scored, and optimized by systems they cannot see or appeal to in any meaningful way. The power asymmetry is stark. The worker is transparent to the algorithm, but the algorithm is opaque to the worker.
Ambient Surveillance and the Internet of Things
Your phone is the most data-rich surveillance device most people carry, but it is no longer the only one. The proliferation of internet-connected devices in homes, cars, and public spaces has created what researchers call ambient surveillance, an environment where data collection happens passively, all the time, without requiring any active input from the person being monitored.
Smart-home research has identified at least 25 types of sensors, including motion sensors, contact sensors, pressure sensors, and electrical current sensors, that can be deployed within rooms, on furniture, or inside appliances in an unobtrusive way.6PubMed Central. Unobtrusive Health Monitoring in Private Spaces: The Smart Home The stated purpose of these sensors is often health monitoring or home automation, helping elderly residents age in place or letting you adjust your thermostat from the couch. But the data these sensors generate, when aggregated, paints an extraordinarily detailed picture of daily life: when you wake up, when you eat, how often you open the fridge, whether you are moving normally or showing signs of a fall, even your bathroom habits.
The surveillance-capitalism question is not whether any single data point matters. It is what happens when all these data points are connected. A smart thermostat company may be benign on its own. But when that company is acquired by a tech conglomerate that also runs an advertising platform, your home temperature data joins your search history, your email content, and your location trail in a single behavioral profile. The data does not stay siloed for long.
The Psychological Cost
Living under pervasive data collection has measurable effects on how people feel and behave, even when they are not consciously thinking about surveillance. A comprehensive review of the relationship between security technologies and mental health found that while these technologies can provide some sense of safety, they also carry negative consequences, including privacy invasion, surveillance-related stress, paranoia, and ethical concerns that can erode mental health.7PubMed Central. Exploring the Impact of Security Technologies on Mental Health: A Comprehensive Review
The psychological effects are not limited to people who are actively worried about privacy. Research in social psychology has repeatedly shown that awareness of being watched changes behavior, a phenomenon that predates digital technology by decades. People become more conformist, less creative, and more risk-averse when they believe they are being observed. Surveillance capitalism extends the observation to nearly every waking moment. You may not be thinking about the fact that your phone tracks which aisle you linger in at the grocery store, but the knowledge that such tracking is possible, once internalized, can produce a low-grade wariness that accumulates over time.
There is also the problem of chilling effects. When people know their communications might be monitored or their browsing history stored indefinitely, some avoid searching for sensitive health information, skip certain political content, or self-censor in messaging apps. The harm is not that anyone gets punished for a particular search. The harm is that the range of free inquiry quietly narrows.
Education Technology and Student Data
Schools and universities have become major sites of data extraction, often without much public attention. The rapid adoption of educational technology platforms, accelerated by the shift to remote learning during the pandemic, has meant that children as young as five now generate substantial digital footprints through their schoolwork. Learning management systems, adaptive testing platforms, and plagiarism detection tools all collect data on how students learn, when they struggle, how quickly they read, and what distracts them.
Scholarly discussion has raised concerns about the tension between big data analytics in education and student privacy, noting that the same tools promising personalized learning also create detailed behavioral profiles of minors.8Sage Journals. Achieving big data privacy in education The problem is compounded by the fact that students, especially younger ones, rarely have meaningful ability to consent to or opt out of these systems. When a school district adopts a particular platform, students either use it or fall behind. The commercial entities behind these platforms gain access to years of longitudinal behavioral data on children, which has obvious value for prediction markets even if the immediate stated purpose is educational improvement.
Parents and educators are often unaware of the scope of this data collection. A reading app that seems to just track whether a child finished a book may also record reading speed, re-reading patterns, time of day, and device usage context. Whether that data is retained, sold, or shared with third parties depends on terms-of-service agreements that few people read and fewer understand.
The Global Dimension
Surveillance capitalism is often discussed as though it were a neutral market phenomenon, but it has a clear geography. The companies that dominate the extraction and sale of behavioral data are overwhelmingly headquartered in a small number of wealthy countries, while the people whose data is extracted live everywhere. This creates a dynamic that some scholars describe as digital colonialism.
Research examining the role of US technology firms in the Global South has proposed a framework for understanding how the United States is reinventing colonial relationships through the domination of digital technology. Using South Africa as a case study, this work argues that US multinationals exercise control at the architecture level of the digital ecosystem, including software, hardware, and network connectivity, which then enables related forms of domination. As a feature of surveillance capitalism, the research notes, intelligence agencies in the Global North partner with their own corporations to conduct mass and targeted surveillance in the Global South, intensifying imperial state surveillance.9Race & Class. Digital colonialism: US empire and the new imperialism in the Global South
In practical terms, this means that a farmer in Kenya using a free mobile platform to sell crops, a student in Brazil using a free email service for university, and a health worker in India entering patient data into a cloud-based system are all feeding behavioral surplus into an extraction pipeline that primarily benefits firms and shareholders in Silicon Valley. The services they receive in return are real and sometimes valuable, but the terms of the exchange are set unilaterally, and the data flows almost entirely in one direction. Countries in the Global South frequently lack the regulatory infrastructure to set their own terms, and when they try, they face pressure from both the firms themselves and from foreign governments with strategic interests in maintaining the data flow.
Why Regulation Has Been Slow
The European Union’s General Data Protection Regulation, which took effect in 2018, remains the most ambitious attempt to regulate the data extraction economy. It established principles like data minimization (collect only what you need), purpose limitation (use data only for the reason you collected it), and explicit consent. Other jurisdictions have followed with their own versions, including California’s consumer privacy act and Brazil’s general data protection law.
But these laws face structural challenges. Enforcement is slow and underfunded relative to the scale of the companies involved. Consent mechanisms have largely devolved into cookie banners and pop-ups that most people click through without reading, making “informed consent” more of a legal fiction than a genuine safeguard. And the fundamental business model, trading free services for behavioral data, is not something any major regulation has yet attempted to outlaw. Regulators have focused on limiting the worst abuses rather than challenging the underlying logic.
There is also a political dimension. Governments themselves benefit from the data infrastructure that surveillance capitalism builds. Intelligence and law enforcement agencies have access to commercial data troves through legal requests, data-sharing agreements, or outright purchases from data brokers. Regulating the industry too aggressively could reduce the state’s own surveillance capacity, which creates a conflict of interest that critics argue has slowed legislative action in many countries.
What Individual Action Can and Cannot Do
Privacy-conscious individuals can take steps to reduce their data footprint: using encrypted messaging apps, browser extensions that block trackers, VPNs, and search engines that do not log queries. These tools are real and effective at the margins. But framing the problem as one of individual consumer choice misses the structural nature of the system. You can choose not to use a particular social network, but you cannot choose to live in a society where your face is not captured by surveillance cameras, your friends do not tag you in photos, or your employer does not require you to use data-harvesting productivity software.
The “privacy as personal responsibility” framing is itself useful to the surveillance-capitalism model, because it shifts the burden from companies that extract data to individuals who are asked to defend against extraction. It is the digital equivalent of telling people to recycle their way out of climate change. The recycling helps, but the factory is the problem. Meaningful change requires structural intervention: stronger regulation, enforcement with real penalties, public investment in non-commercial digital infrastructure, and a willingness to question whether free services funded by behavioral data extraction are actually a good deal for society.
The gap between what individuals can do and what the system requires is where much of the frustration around surveillance capitalism lives. People sense that something is off about the exchange, that the services they receive are not worth the intimacy of the data they surrender, but the alternatives are either inconvenient, expensive, or simply do not exist yet for many daily tasks. Until the structural incentives change, the extraction will continue to outpace the resistance.