01: THE ENEMY

Government surveillance comes in many forms. We will focus on a few of the most prevalent forms here. Government surveillance can be sorted into four main categories: public surveillance (think facial recognition cameras and license plate readers), interception (i.e. wiretapping and scanning of communications and internet traffic), device searches, and data purchasing (when the government buys information from data brokers). The most relevant categories to this article are the first two, though all four are dangerous and quite prevalent in the US. 

PUBLIC SURVEILLANCE

Public surveillance is perhaps the most concerning. In recent years, the most problematic forms of this have been facial recognition and algorithm-based screening. According to the Electronic Frontier Foundation (EFF), many school districts have implemented these technologies to screen students for "aggression" in real time, often coming up with false positives (and rarely with true ones). In addition to being an Orwellian from of surveillance, these tools disproportionately affect marginalized and over-policed groups due to bias in their algorithms and their implementation primarily in communities with high populations of people of color.

These systems work in a few ways. A facial recognition camera (in the context of live surveillance, not the ones at TSA) is basically just a normal security camera connected to a computer running a face detection algorithm. Depending on the purpose, it can do different things with the data, but usually it runs it through a database of faces to determine who it's seeing. The computer then reports this information back to the camera's operator. This kind of system is commonly used by police to identify the perpetrators of crimes and make arrests, though the reliability of these systems has been called into question and they have led to numerous false arrests. 

Facial recognition has been used in other places as well. One example of this kind of facial recognition is automated passport control at airports, which use facial recognition to verify that the person trying to enter the country is the same one on the passport. While on paper this type of screening seems innocent, it still has real harms in practice. By taking photos associated with identifying documents such as a boarding pass or passport, the government (and in some cases the airlines as well) can create a database of faces, which can be used for more sinister live surveillance. These databases, like all data stored on the cloud, are also vulnerable to data breaches, which could put the biometric information of thousands of people into the hands of threat actors. Then there is the issue of informed consent. Having to seek out an agent and request an exception places a significant burden on travelers wishing to opt out of the system, and some don't even know they can. 

Another type of public surveillance that has gained attention in recent years are automated license plate readers (ALPRs). The name is pretty self explanatory; they are automatic devices that read license plates. Often used at tolls on highways, these cameras are used to identify people driving on the road. Digital tolls aren't a problem, but in some places, these cameras have spread onto city streets, which has caused issues for people driving to abortion clinics after the Supreme Court overturned Roe v. Wade. 

When these technologies are combined with location data purchased from data brokers, they can be used by law enforcement to track where any given person is at any given time. This level of surveillance begins to sound like that of The Party in George Orwell's 1984. 

INTERCEPTION

But wait! We still haven't run out of Big Brother comparisons. And who better to use them on than the NSA? When thinking of interception surveillance, these three letters come to mind instantly. The agency exists almost solely for the purposes of interception, usually of the communications of innocent Americans without any form of warrant (or any oversight at all for that matter). They even break the (already constitutionally shaky) laws that gave them their power in the first place, according to the Washington Post. Their interception surveillance works pretty intuitively: The NSA puts interceptors at communication hubs (like cell towers and undersea cables) that collect internet traffic and send it back to them. Then they collect the traffic in a database which may or may not be stored in a $2 billion data center in Utah. What they use it for is anyone's guess, as the whole operation is top secret.

THE VIRTUAL WALL

You may have heard of the wall: Trump's attempt to build a fence along the whole US-Mexico border. But there is a lesser known network on the border that is much more extensive. The virtual border wall consists of surveillance towers, drones, blimps, ALPRs, and other technologies in an attempt to identify people sneaking into the US and prevent drug trafficking. The system exists on both sides of the border, and the Mexican wall might even be more extensive than the American one. In addition to allegations that the incredibly costly system (over 1 billion dollars since 2005 according to MIT Technology Review) just doesn't work, it could also lead to massive human rights abuses in communities near the border as people's every move is catalogued by thousands of cameras and other sensors.

THE ARREST OF RANDALL QURAN REID

In November of 2022, Randall Quran Reid, a 29-year-old Black man from Georgia, was arrested for buying two bags worth $8000 from a shop in Luisiana with a stolen credit card. The only evidence against him was a facial recognition match from a surveillance video. The police didn't bother to get any more evidence—which would have revealed that Reid was in Georgia when the crime was committed—before arresting him. Unfortunately, his story is very common, as facial recognition systems are notoriously unreliable, especially when identifying people of color.

NSA COURT CASES

Many organizations have sued the NSA over their surveillance of Americans not suspected of crimes, but all of these cases have been dismissed due to the "state secrets defense". This basically means the NSA can do whatever they want because the evidence against them is all classified, and apparently knowledge of how the government spies on its own innocent citizens could compromise national security.

PREDICTIVE POLICING

No, its not some crazy thing out of Minority Report. Predictive policing is a real technology that uses records of past criminal activity to predict where police units are needed most or who is most likely to commit a crime. This technology is often used in conjunction with surveillance systems, as cities that have the ability to implement it usually have technologically advanced police forces already. Surveillance tech like facial recognition cameras are often used to gather data for these systems or track individuals deemed high-risk. The problem with these systems is that they create feedback loops: if the police arrest a lot of people in an area, the system will direct more police to that area, and more arrests will be made. This can have the effect of putting more police into already overpoliced communities while moving them away from areas where less crimes are reported, though there could still be criminal activity. As a result, many cities have already banned predictive policing or placed restrictions on its use.

Ciudad Juarez is the largest city in the Mexican state of Chihuahua. Located on the US-Mexico border, the city attracts many migrants seeking entry into the US, as well as gangs smuggling drugs and people into the US. Due to the city's high crime rate (110 homicides in May 2023 alone, according to InSight Crime), the government of Chihuahua began building a massive surveillance network there in 2022 called the Plataforma Centinela. Consisting of over 9,600 cameras (including 1700 ALPRs) equipped with facial recognition, identification arches with biometric sensors to identify people and vehicles entering and leaving the state, and a fleet of 74 surveillance drones, the network aims to monitor all activity in Juarez and 13 other municipalities in the state of Chihuahua in real time. All of the data will be analyzed in real time by artificial intelligence, which suggests the use of predictive policing technology, though the government has been very secretive about the project. The government of Chihuahua has signed an agreement with Texas to allow law enforcement on the US side of the border access to the Plataforma Centinela for use in border control.

This approach of surveillance is certainly not without issues. Yes, there is a lot of crime, and yes, arresting criminals does decrease crime. But it is not a long term solution. In the case of Juarez, police surveillance has a higher budget than social services, which actually address the underlying issues causing the city's alarming crime rate. All told, the Plataforma Centinela will cost the government of Chihuahua 4.2 billion pesos, plus additional operation and maintenance costs. (The state also will not own the system until they finish paying off the costs of construction, under a contract which violates public works law according to Norte Digital.) Meanwhile, the budget for social services is around 40 million pesos. In a city where about 60% of the population lives in some degree of poverty (according to Statista), this ammount of public money being directed away from social services to fund surveillance could have significant negative effects and slow social and economic development in the city. 

In addition, the system violates residents' human right to privacy, established under the International Covenant on Civil and Political Rights (ICCPR), by subjecting them to constant surveillance. This surveillance also has the ability to chill free speech and protest, as residents know their activities are being recorded at all times. 

Then there is the risk of wrongful arrests. The Plataforma Centinela relies on algorithms and AI to analyze the data it collects in real time, including determining the location and perpetrator of a crime. The system makes use of facial recognition and predictive policing, both of which have high false positive rates. These technologies have led to many wrongful arrests, leading several cities to ban them. Police responding to these false positive alerts risk using force on innocent people, and false alarms draw police resources away from actual crimes. Combined with predictive policing algorithms that use alerts as part of their calculations, the system creates a feedback loop where every false alarm in an area increases the chance of another one, meaning some areas could be subject to constant police presence when no crimes have been committed.

Surveillance systems do have benefits. They can help police catch criminals faster and solve cases that otherwise would go cold. The problem arises when the risks outweigh the benefits. For example, a simple CCTV system in a building is generally not considered problematic, because it increases security without greatly compromising the privacy of the occupants or harming innocent people. But if the same building had biometric cameras and microphones covering every room, the privacy impact would be exponentially greater, while safety and security would not see similar increases.

Evaluating the benefits and risks of surveillance tools is complicated, and depends on the type(s) of surveillance used, the scale, the use case, and other factors specific to each deployment.

Common arguments in favor of surveillance technology are that it can detect and deter crime, help police in investigations, and be used in disaster response. Arguments against surveillance include the threat to privacy, wrongful arrests and dangerous encounters with police expecting to find criminals, and threats to freedom of movement and the right to protest.

While some amount of surveillance is necessary to maintain public safety, the level of surveillance found in many modern cities has gone far beyond what is necessary, infringing on the right of residents to live their lives without excessive interference from the government. The risks of these systems far outweigh their benefits, and even these benefits have been called into question by the alarming rate of wrongful arrests they have caused, and the rate of actual crimes they have missed. In some cases, crime even increased after the implementation of a surveillance system. With interception surveillance, the case is even more clear-cut: the amount of data collected under programs such as Section 702 is so great that it is basically useless, and most of it belongs to innocent civilians who aren't even suspected of a crime. This type of surveillance is nothing but a violation of our rights, and it needs to go.

• EFF: Scholars Under Surveillance: How Campus Police Use High Tech to Spy on Students

• EFF: Atlas of Surveillance

• EFF: Street Level Surveillance

• Brookings: Police surveillance and facial recognition: Why data privacy is imperative for communities of color

• Brennan Center: Social Media Surveillance by the U.S. Government

• NPR: Surveillance And Local Police: How Technology Is Evolving Faster Than Regulation

• Time: How the Surveillance of Immigrants Remade American Policing

• ACLU: Community Control Over Police Surveillance: Technology 101