The Surveillance Machine

Harvard Business School professor Shoshana Zuboff coined the term "surveillance capitalism" in 2014 to describe an economic logic that had quietly become the dominant business model of the digital economy. The logic: human experience is raw material. The product is behavioral prediction. The customer is the advertiser (and, increasingly, the state). The user is not the customer — the user is the inventory.

Google and Meta (Facebook) created this model. They provide services — search, social networking, email, mapping — without charging users. Instead, they harvest data about every action, every preference, every location, every relationship, every uncertainty, and every desire. This data is processed into "prediction products": commodities that tell advertisers, political campaigns, and governments what individual users are likely to do, buy, believe, or vote for.

The scale is difficult to comprehend: Google processes over 8.5 billion searches per day. Meta's advertising system tracks users across millions of websites that don't belong to Meta, through invisible tracking pixels embedded by publishers. The combined surveillance infrastructure of these two companies is larger than any state surveillance system in human history — and it is owned by private actors with no democratic accountability.

In 2018, it was revealed that Cambridge Analytica — a data analytics firm funded by billionaire Robert Mercer and connected to Steve Bannon — had harvested the personal data of approximately 87 million Facebook users without their consent. Facebook had allowed third-party developers to collect data not only on app users but on their entire social network. Cambridge Analytica used this data to build psychological profiles and deliver targeted political advertising in the 2016 US election and the Brexit referendum.

The FTC fined Facebook $5 billion for the Cambridge Analytica violations — the largest US privacy fine in history at the time. Zuckerberg personally signed a consent decree in 2012 promising not to share user data without consent; the FTC found he had violated it.

Cambridge Analytica was not an aberration. It was the use case that surveillance capitalism was always structurally available for. The data exists. The targeting exists. The only question is who has access to it and what they do with it. Cambridge Analytica answered that question: they used it to run influence operations on democratic elections.

On August 5, 2024, US District Judge Amit Mehta ruled that Google illegally maintained monopolies in the general search services and general search text advertising markets, in violation of Section 2 of the Sherman Antitrust Act. The ruling came after a trial brought by the Department of Justice and state attorneys general.

Google controls approximately 90% of the US general search market and even higher shares of mobile search. The court found that Google maintained this dominance through exclusive agreements — paying Apple approximately $18–20 billion per year to be the default search engine on Safari — that effectively locked out competitors regardless of quality.

The remedies phase of the case continued into 2025, with the DOJ seeking structural remedies including possible divestiture of Chrome or Android. The monopoly finding itself was not appealed — Google's legal position acknowledged the market share but contested the method. The ruling confirmed what researchers had documented for years: Google's dominance is not the product of superior search alone, but of systematic exclusionary practices.

Meta's advertising infrastructure extends far beyond its own platforms. Through the Meta Pixel — a small piece of JavaScript code embedded on millions of external websites — Meta tracks user behavior across the broader web: what articles users read, what products they look at, what health conditions they search for, what legal services they research. This data is collected regardless of whether the user is logged in to Facebook.

ProPublica and The Markup have documented specific harms from this tracking: in 2022, The Markup found that the Meta Pixel was embedded on tax preparation sites (TurboTax, H&R Block, TaxAct) and transmitting users' income and tax refund data to Meta. In 2022, hospitals were found to have transmitted patient data — including appointment types and medication information — to Meta through embedded Pixels on patient portals.

These are not bugs. They are the feature operating as designed in a system where comprehensive data collection is the revenue model, and where the people whose data is collected have no meaningful recourse or consent mechanism.

The NSA's PRISM program, revealed by Edward Snowden in 2013, showed that the US government had direct access to data held by Google, Apple, Facebook, Microsoft, and Yahoo. But government access to private surveillance data extends beyond NSA programs: federal and state law enforcement agencies routinely purchase data from data brokers — companies that aggregate information from social media, mobile apps, retail loyalty programs, and other sources — without warrants.

A 2023 Office of the Director of National Intelligence report acknowledged that intelligence agencies had purchased "commercially available information" including location data, internet activity, and financial data from commercial data brokers. The report acknowledged the practice raised "significant" civil liberties concerns.

The surveillance capitalism infrastructure is, in practice, a dual-use system: the same data profiles that tell advertisers which users might buy a car or switch credit cards tell law enforcement which users attended a political rally, visited an abortion clinic, or associated with an organizing campaign. The distinction between commercial and government surveillance has been effectively collapsed.