The Government Is Watching You Right Now and Here Is Exactly How
Curiosity
Download the Curiosity App for discussion, debates and more for free.

You are being tracked right now. Not by a shadowy hacker. By a system your own government built, funded, and is actively expanding.

The part everyone skips: how the data actually flows

Most people imagine surveillance as a camera on a pole. That is the kindergarten version. The real system is a layered pipeline that starts the moment you open an app, tap a map, or walk past a license plate reader. Every one of those actions generates a data point. That data point gets sold.

Here is the core mechanic: data brokers amass your web searches, financial records, and location history from millions of people and sell them to various clients, including the U.S. government. Your route to work, your late night searches, your protest attendance: all of it is sitting on hard drives around the world, waiting to be used.

I remember the first time I read about the "real-time bidding" ad system and felt genuinely sick. Every time you open a website or app, information about you is broadcast within a millisecond to thousands of advertisers bidding to put an ad in front of you. Surveillance vendors figured out they could tap that same marketplace to track people without a warrant.

While companies can manipulate you, they cannot put you in jail. But the U.S. government can, and it now purchases massive quantities of your information from commercial data brokers.

The Conversation / Fast Company, April 2026

Step by step: what happens to your data after it leaves your phone

Step one is collection. Your phone pings cell towers constantly. Your doorbell camera feeds a private network. Your neighborhood's Flock license plate readers log every car that passes. None of this requires a warrant because it happens in "public."

Step two is aggregation. Companies like Palantir and Babel Street pull these streams together into massive government databases. Palantir's ICE app, called ELITE, populates a map with potential targets, builds a dossier on each person, and generates a confidence score on their current address.

CCTV surveillance cameras mounted on a city wall, representing the physical layer of the modern government surveillance network.

Step three is AI analysis. This is where it gets genuinely frightening. LLM agents could potentially do the work of intelligence analysts in a fraction of the time and for a fraction of the cost, enabling the state to aim its eye at anyone, not just high priority targets.

The automated nature of these systems means traditional limits on police resources have been effectively eliminated. The machine never sleeps, never forgets, and never needs a judge to sign off.

The legal fiction that makes all of this possible

Here is the trick the government uses to avoid the Fourth Amendment. The Constitution prohibits unreasonable search and seizure. Supreme Court cases require a warrant to track someone's phone location. So the government simply buys the data instead of seizing it.

On March 18, 2026, FBI Director Kash Patel confirmed to Congress that the FBI is buying Americans' data from data brokers, including location histories, to track citizens. He said it out loud. In public. To Congress. And almost nothing happened.

The government's argument is this: you voluntarily gave your data to an app, so you have no privacy expectation in it. I find this argument completely unserious. Nobody "volunteers" their location to a Muslim prayer app expecting the Pentagon to buy it and run it through an AI profiling engine.

The funding surge that should alarm everyone

This is not a fringe operation running on a shoestring. The 2025 spending law gave DHS an unprecedented $165 billion in yearly funding, with ICE alone receiving roughly $86 billion. DHS also put a $1 billion ceiling on a single contract with Palantir that went into effect in February 2026.

DHS is reportedly funding companies to convert agents' phones into biometric scanners and build an AI platform that acquires all 911 call center data to create geospatial heat maps predicting incident trends. That last one should stop you cold.

Data center server racks glowing blue, representing the AI infrastructure behind modern government surveillance systems.

The strongest counterargument, and why it does not hold up

The national security case for all of this is real. Facial recognition has helped find missing children. Location data has broken up trafficking rings. I am not pretending the tools have zero legitimate use.

But here is the problem: seven federal law enforcement agencies used facial recognition services without requiring any staff training, conducting roughly 60,000 searches before a single training requirement existed. Only two of those seven agencies had implemented training as of April 2023. This is not a system with guardrails. It is a system with a budget.

A 2016 Georgetown Law study found that half of all American adults' faces were already in law enforcement facial recognition databases. You did not consent to that. You were not asked.

The technology allows basically a panopticon. You can just have AI finding the patterns, aggregating data and allowing the government to build this enormous surveillance state that threatens our civil liberties.

Brendan Steinhauser, CEO, Alliance for Secure AI, via NBC News

What you can actually do, and what will not save you

Deleting apps helps at the margins. Using a VPN helps a little more. But the honest answer is that individual privacy hygiene cannot outrun a $165 billion institutional machine.

The only real lever is political. The Electronic Frontier Foundation and the ACLU are fighting in courts and legislatures right now. Section 702 of FISA, the legal authority that enables much of this warrantless collection, is up for reauthorization and the fight over it is the most important privacy battle most people have never heard of.

Would you trust a system that ran 60,000 warrantless facial recognition searches before anyone was trained to use it? Because that system is already running. On you.