As technology becomes more and more integrated into our daily lives, we are willfully giving away our data to private companies. The issue lies in the fact that most users don't actually know where this data is going, and how it is being used.

86%

of Americans willing to discuss government surveillance in person.

42%

willing to post about it online.

Pew Research Center, 2014

As technology becomes more and more integrated into our daily lives, we are willfully giving away our data to private companies. The issue lies in the fact that most users don't actually know where this data is going, and how it is being used.

86%

of Americans were willing to discuss government surveillance in person.

42%

were willing to post about it online.

Pew Research Center, 2014

As technology becomes more and more integrated into our daily lives, we are willfully giving away our data to private companies. The issue lies in the fact that most users don't actually know where this data is going, and how it is being used.

86%

of Americans were willing to discuss government surveillance in person.

42%

were willing to post about it online.

Pew Research Center, 2014

Most people would assume that their data lives only within the app they gave it to.

Most people would assume that their data lives only within the app they gave it to.

Unfortunately, it does not.

Unfortunately, it does not.

Most people would assume that their data lives only within the app they gave it to.

Most people would assume that their data lives only within the app they gave it to.

Unfortunately, it does not.

Unfortunately, it does not.

Private companies collect, aggregate, and sell your data, often to other private companies to push targeted ads and products to you, but also government agencies. And because most of this data is being bought and not subpoenaed, the legal protections that would normally apply to private data don't here.

Private companies collect, aggregate, and sell your data, often to other private companies to push targeted ads and products to you, but also government agencies. And because most of this data is being bought and not subpoenaed, the legal protections that would normally apply to private data don't here.

Before the rise of AI, this massive amount of data would be impossible to comb through with only humans, which is why surveillance was typically a targeted operation. But now, AI has given us the capability to sort all this data into a detailed profile of who you are and what you do on a day to day. What seemed not too long ago like a dystopian concept, is now being used inside the US, by agencies and contractors most people have never even heard the name of.

Before the rise of AI, this massive amount of data would be impossible to comb through with only humans, which is why surveillance was typically a targeted operation. But now, AI has given us the capability to sort all this data into a detailed profile of who you are and what you do on a day to day. What seemed not too long ago like a dystopian concept, is now being used inside the US, by agencies and contractors most people have never even heard the name of.

86%

86%

of Americans were willing to discuss government surveillance in person, while only

of Americans were willing to discuss government surveillance in person, while only

42%

42%

were willing to post about it online.

were willing to post about it online.

Pew Research Center, 2014

This gap is the outcome surveillance was always designed to produce. Surveillance doesn't have to be actively used to shape behavior. The awareness of it is enough, but that awareness didn't emerge on its own. It is the product of decades of surveillance infrastructure quietly expanding and embedding itself into everyday life.

This gap is the outcome surveillance was always designed to produce. Surveillance doesn't have to be actively used to shape behavior. The awareness of it is enough, but that awareness didn't emerge on its own. It is the product of decades of surveillance infrastructure quietly expanding and embedding itself into everyday life.

This gap is the outcome surveillance was always designed to produce. Surveillance doesn't have to be actively used to shape behavior. The awareness of it is enough, and that is the product of decades of surveillance infrastructure quietly expanding and embedding itself into everyday life.

This gap is the outcome surveillance was always designed to produce. Surveillance doesn't have to be actively used to shape behavior. The awareness of it is enough, and that is the product of decades of surveillance infrastructure quietly expanding and embedding itself into everyday life.

1908

The FBI is founded. Surveillance of political dissidents and civil rights organizers begins almost immediately.

1908

The FBI is founded. Surveillance of political dissidents and civil rights organizers begins almost immediately.

1956

COINTELPRO launches. The federal government systematically monitors and disrupts civil rights organizations, anti-war groups, and Black activists.

1956

COINTELPRO launches. The federal government systematically monitors and disrupts civil rights organizations, anti-war groups, and Black activists.

2001

The Patriot Act passes 45 days after 9/11. It expands government authority to monitor the communications and financial records of ordinary citizens without a warrant.

2001

The Patriot Act passes 45 days after 9/11. It expands government authority to monitor the communications and financial records of ordinary citizens without a warrant.

2013

Edward Snowden reveals the NSA's PRISM program. Mass collection of internet communications from millions of Americans had been occurring with no individual suspicion required.

2013

Edward Snowden reveals the NSA's PRISM program. Mass collection of internet communications from millions of Americans had been occurring with no individual suspicion required.

2015

The Department of Homeland Security begins monitoring Black Lives Matter on social media. Protest activity is tracked in real time as demonstrations spread across the country.

2015

The Department of Homeland Security begins monitoring Black Lives Matter on social media. Protest activity is tracked in real time as demonstrations spread across the country.

2018

Cambridge Analytica becomes public. The personal data of 87 million Facebook users had been harvested without consent and used for political targeting.

2018

Cambridge Analytica becomes public. The personal data of 87 million Facebook users had been harvested without consent and used for political targeting.

2020

Federal agencies compile surveillance files on racial justice protesters and deploy aerial surveillance over American cities. Most of those arrested during protests were charged with nonviolent misdemeanors.

2020

Federal agencies compile surveillance files on racial justice protesters and deploy aerial surveillance over American cities. Most of those arrested during protests were charged with nonviolent misdemeanors.

2025

AI systems now automate what once required hundreds of agents. A single platform can cross-reference license plates, social media, financial records, and biometric data in seconds.

2025

AI systems now automate what once required hundreds of agents. A single platform can cross-reference license plates, social media, financial records, and biometric data in seconds.

How to navigate
Hover over any node to see details. Scroll to zoom, click and drag to pan. Use the + / − buttons to zoom in and out.

These systems do not affect everyone equally though, and surveillance systems don't create bias from scratch. They inherit it. Predictive policing algorithms are trained on historical arrest data, which is shaped by decades of discriminatory enforcement.

These systems do not affect everyone equally though, and surveillance systems don't create bias from scratch. They inherit it. Predictive policing algorithms are trained on historical arrest data, which is shaped by decades of discriminatory enforcement.

The result is a feedback loop: over-policed communities generate more data, which the algorithm interprets as higher risk, which directs more policing toward those same communities. The result is a system that doesn't need to be explicitly racist to produce racist outcomes. The data does that work on its own.

The result is a feedback loop: over-policed communities generate more data, which the algorithm interprets as higher risk, which directs more policing toward those same communities. The result is a system that doesn't need to be explicitly racist to produce racist outcomes. The data does that work on its own.

These systems do not affect everyone equally though, and surveillance systems don't create bias from scratch. They inherit it. Predictive policing algorithms are trained on historical arrest data, which is shaped by decades of discriminatory enforcement.

These systems do not affect everyone equally though, and surveillance systems don't create bias from scratch. They inherit it. Predictive policing algorithms are trained on historical arrest data, which is shaped by decades of discriminatory enforcement.

The result is a feedback loop: over-policed communities generate more data, which the algorithm interprets as higher risk, which directs more policing toward those same communities. The result is a system that doesn't need to be explicitly racist to produce racist outcomes. The data does that work on its own.

The result is a feedback loop: over-policed communities generate more data, which the algorithm interprets as higher risk, which directs more policing toward those same communities. The result is a system that doesn't need to be explicitly racist to produce racist outcomes. The data does that work on its own.

77%

77%

higher chance of being flagged as high risk by a predictive policing algorithm if you are Black — even with a similar record to a white defendant.

higher chance of being flagged as high risk by a predictive policing algorithm if you are Black — even with a similar record to a white defendant.

ProPublica, 2016

700,000+

700,000+

pages of federal surveillance documents obtained through FOIA litigation showing agencies monitored racial justice protesters and labeled them threats.

pages of federal surveillance documents obtained through FOIA litigation showing agencies monitored racial justice protesters and labeled them threats.

Brennan Center for Justice, 2022

17,000+

17,000+

people arrested across 50 cities in the first two weeks of BLM protests in 2020. Most were charged with nonviolent misdemeanors like curfew violations.

people arrested across 50 cities in the first two weeks of BLM protests in 2020. Most were charged with nonviolent misdemeanors like curfew violations.

CBS News / DHS records

77%

77%

higher chance of being flagged as high risk by a predictive policing algorithm if you are Black — even with a similar record to a white defendant.

higher chance of being flagged as high risk by a predictive policing algorithm if you are Black — even with a similar record to a white defendant.

ProPublica, 2016

700,000+

700,000+

pages of federal surveillance documents obtained through FOIA litigation showing agencies monitored racial justice protesters and labeled them threats.

pages of federal surveillance documents obtained through FOIA litigation showing agencies monitored racial justice protesters and labeled them threats.

Brennan Center for Justice, 2022

17,000+

17,000+

people arrested across 50 cities in the first two weeks of BLM protests in 2020. Most were charged with nonviolent misdemeanors like curfew violations.

people arrested across 50 cities in the first two weeks of BLM protests in 2020. Most were charged with nonviolent misdemeanors like curfew violations.

CBS News / DHS records

77%

77%

higher chance of being flagged as high risk by a predictive policing algorithm if you are Black — even with a similar record to a white defendant.

higher chance of being flagged as high risk by a predictive policing algorithm if you are Black — even with a similar record to a white defendant.

ProPublica, 2016

700,000+

700,000+

pages of federal surveillance documents obtained through FOIA litigation showing agencies monitored racial justice protesters and labeled them threats.

pages of federal surveillance documents obtained through FOIA litigation showing agencies monitored racial justice protesters and labeled them threats.

Brennan Center for Justice, 2022

17,000+

17,000+

people arrested across 50 cities in the first two weeks of BLM protests in 2020. Most were charged with nonviolent misdemeanors like curfew violations.

people arrested across 50 cities in the first two weeks of BLM protests in 2020. Most were charged with nonviolent misdemeanors like curfew violations.

Washington Post, 2020

These are not edge cases or anomalies. They are the predictable outputs of systems built on data that was never neutral to begin with.

These are not edge cases or anomalies. They are the predictable outputs of systems built on data that was never neutral to begin with.

When challenged, these systems are consistently defended with the same justifications. Below are some of the most common claims made in defense of surveillance infrastructure — and what the documented evidence says about each of them.

When challenged, these systems are consistently defended with the same justifications. Below are some of the most common claims made in defense of surveillance infrastructure — and what the documented evidence says about each of them.

What you're told

License plate readers are only used to solve serious crimes.

tap to reveal
What the data shows

A Ventura County Sheriff audit found that outside agencies accessed one county's plate data 364,000 times in a single month. 299 of those searches cited immigration enforcement. The city of Woodburn, OR discovered CBP had accessed their Flock Safety network 384 times through a pilot program they were never informed about.

tap to flip back
What you're told

License plate readers are only used to solve serious crimes.

tap to reveal
What the data shows

A Ventura County Sheriff audit found that outside agencies accessed one county's plate data 364,000 times in a single month. 299 of those searches cited immigration enforcement. The city of Woodburn, OR discovered CBP had accessed their Flock Safety network 384 times through a pilot program they were never informed about.

tap to flip back
What you're told

License plate readers are only used to solve serious crimes.

tap to reveal
What the data shows

A Ventura County Sheriff audit found that outside agencies accessed one county's plate data 364,000 times in a single month. 299 of those searches cited immigration enforcement. The city of Woodburn, OR discovered CBP had accessed their Flock Safety network 384 times through a pilot program they were never informed about.

tap to flip back

The gap between official justifications for these systems and what the data actually shows is consistent. Below are some of the most common claims made in defense of surveillance infrastructure — and what the documented evidence says about each of them.

The gap between official justifications for these systems and what the data actually shows is consistent. Below are some of the most common claims made in defense of surveillance infrastructure — and what the documented evidence says about each of them.