The first principle of the data economy was once "free service in exchange for advertising." But the surveillance capitalism business model is completely different: revenue comes from prediction markets; profit comes from "behavior prediction multiplied by targeting algorithms."
Shoshana Zuboff described this mechanism precisely in her 2022 testimony to the US Congress: "Clandestine observation methods secretly extract what was once considered private human experience and convert it into behavioral data. These methods operate outside human consciousness, depriving the actor of the right to know, and the right to resist that follows."
Zuboff breaks down the operation of surveillance capitalism into four steps: data extraction → behavioral datafication → predictive computation → sale of predictions. She cites leaked 2018 Facebook documents: the company's "prediction engine" "ingests trillions of data points per day," produces thousands of "models," and its "prediction services" produce "over 6 million predictions per second."
This is not the business of "data." This is the business of "human behavior futures." Zuboff compares it to markets for "pork futures, wheat, or oil futures" — except the traded object is "human future behavior."
If the business model of the data economy is "behavior prediction," then its structural cost is the systematic erosion of cognitive capacity. A longitudinal study published in the Journal of Adolescent Health in September 2026 (N=9,747, from the ABCD study) reveals this mechanism:
Cognitive control is the mediator. When adolescents increased social media browsing at baseline, their cognitive control declined two years later (β = −.06, p = .020), which in turn predicted increased depressive symptoms four years later (β = −.06, p = .010).
This is not a problem of "using it too long." This is a problem of "the cognitive pathway being altered." The research team states explicitly: "Reduced cognitive control may be a general mechanism linking adolescent screen use and social media browsing to later depressive symptoms."
American Enterprise Institute survey data from 2026 further quantifies this structural risk: nearly 30% (29%) of adolescent girls report spending 4 hours or more on social media daily. A larger study (N=14,726, Oxford University) found that "exposure to self-harm content, seeking help online, and posting content later regretted" were associated with higher anxiety, depression, and loneliness scores — while "time" itself was only weakly correlated.
The algorithm is not "providing content." It is "shaping the cognitive environment."
Attributing addiction to "users' lack of self-control" is the most typical form of cost transfer.
A study published at ACM 2026 systematically catalogued the Engagement-Prolonging Designs that "very large online platforms" (VLOPs) deploy for adolescent users. Researchers created European adolescent accounts and conducted content analysis of 16 VLOP apps and 17 websites, identifying four strategies:
- Luring engagement: using pleasure, curiosity, or reward-driven mechanisms
- Pressuring engagement: creating social obligations or pressure
- Hypnotic continuation: using low-friction, autonomy-reducing designs
- Trapping users: setting exit barriers, using overwhelming or confusing designs
Platforms most commonly use combined strategies. The study identified 14 patterns and built a video dataset with 583 interactive instances.
Research on "digital addiction and responsiveness to dark patterns" further confirms: there is a strong association between responsiveness to notification-based dark patterns and digital addiction, robust even after controlling for exposure frequency. The researchers note these findings indicate "the user's responsiveness, rather than mere exposure, is strongly associated with digital addiction."
It is not that "users lack self-discipline." It is that the interface is designed to make self-discipline impossible.
The 2018 Cambridge Analytica scandal was not "a violation by a single company." It was the inevitable product of the structural logic of surveillance capitalism.
Facebook gave user data to third-party apps, which in turn passed it to Cambridge Analytica for voter profiling and targeting. In December 2019, the US Federal Trade Commission issued an opinion finding that Cambridge Analytica "engaged in deceptive conduct to collect personal information from tens of millions of Facebook users for voter profiling and targeting." The FTC ultimately ordered Cambridge Analytica to delete the illegally collected data.
But Cambridge Analytica was only "the tip of the iceberg." Yale Law School research notes that in pursuit of profit, Facebook "granted end-user information access to numerous for-profit companies without adequate safeguards that these companies were not manipulating their end users." This created a conflict of interest — "giving Facebook an incentive to look the other way, which it apparently did."
When data extraction reaches sufficient scale, "manipulation" is not the exception. It is the business model.
The cost transfer of the data economy occurs not only at the individual level but also at the global level.
Research on "global algorithmic governance" notes that global institutions, digital superpowers, and large technology companies are "constructing a global surveillance architecture that shapes people's minds and regulates their daily lives through mass surveillance, monitoring, extraction, and regulation."
These architectures are "often promoted on the grounds that enhanced surveillance brings greater security," but one global digital public infrastructure expert warns: "If only 3-4 major players control most of the world's data, then it's essentially a honey pot."
The Global South contributes data, but the value is captured by Global North companies. Research by the Brazilian Institute of Geography and Statistics notes that Big Tech's "soft power" operations in the official statistics field — data festivals, hackathons, exchange programs — are "part of a gradual strategy to win over national and internal statistical agencies against the resistance advanced by private companies."
This is not "technical cooperation." This is systematic erosion of data sovereignty.
| Dimension | Pharma | Food | Healthcare | Insurance | Education | Aging | Data |
|---|---|---|---|---|---|---|---|
| Lifelong customer | Lifelong medication | Annual seed purchase | Repeated visits | Lifelong premiums | Lifelong loan repayment | Lifelong management fees | Lifelong attention |
| Structural incentive | Prefer chronic | Prefer dependency | Prefer more | Prefer denial | Prefer high tuition | Prefer assets under management | Prefer maximum engagement |
| Cost transfer mechanism | Patent monopoly | Seed patents | Fee-for-service | Denial, delay | Student loans | Risk privatization | Behavioral datafication |
| Financialization | REITs, private equity | Agri-finance complex | Hospital chains | Insurer buybacks | SLABS securitization | Pension asset securitization | Platform equity |
| Structural output | 400,000 deaths | 300,000 farmer suicides | 30% spending unnecessary | 19% denial rate | USD 1.86 trillion debt | 380 million uncovered | Cognitive capacity erosion |
These are not seven systems. This is one system, with seven faces.
"Costs must not be transferred to the voiceless vulnerable."
The cost transfer chain of the data system:
- Users: contribute attention, behavior, psychological data in exchange for "free" services.
- Adolescents: cognitive control systematically impaired, depressive symptoms increase.
- Global South: contributes data, but value is captured by Global North companies.
- Future generations: inherit a cognitive environment shaped by algorithms.
This is not a "market outcome." This is institutional design.
The Structural Diagnosis
Suffering is not a "side effect." Suffering is the product itself.
The machine must end. Not because of moral awakening, but because of arithmetic constraints. When adolescents spend 4+ hours daily on social media, when cognitive control is systematically eroded, when data colonialism turns the Global South into a "honey pot," the system's operating cost has exceeded its return.
The end is not "fixing" it. The end is replacing the incentive structure that produces it: from "higher engagement, higher profit" to "more protected cognitive autonomy, healthier society."