John Oliver Exposes the Rapid Expansion of Police Surveillance Technology Across America
American police departments are rapidly building surveillance networks capable of tracking vehicles, accessing private security cameras, monitoring cellphones, detecting suspected gunfire and using algorithms to predict where crimes may occur. On HBO’s Last Week Tonight, John Oliver examined how these technologies have quietly spread through local law enforcement agencies, often faster than the laws and oversight systems designed to regulate them. The central issue is not whether surveillance technology can help police solve crimes, in certain circumstances, it clearly can. The larger question is what happens when extraordinarily powerful systems capable of monitoring thousands of innocent people become routine pieces of municipal infrastructure.
Real Time Crime Centers Are Becoming Centralized Surveillance Hubs
At the center of this transformation are Real Time Crime Centers, centralized facilities where police departments can combine information from public cameras, private security systems, license plate readers and other surveillance technologies. Platforms such as Fusus allow participating cameras to be integrated into law enforcement networks, potentially giving officers immediate access to video from businesses and other private properties.
Oliver highlighted Rialto, California, where the relationship between local government and the surveillance network has gone considerably further. City requirements have incorporated Fusus compatible security systems into certain business permitting and occupancy processes, effectively expanding the number of privately owned cameras that can become part of the police surveillance infrastructure.
The technology presents an obvious public safety argument. During an active shooting, kidnapping, robbery or search for a dangerous suspect, immediate access to nearby cameras could provide officers with critical information. But the same infrastructure does not disappear after an emergency ends. Once thousands of cameras are connected to centralized systems, police potentially possess a powerful network capable of observing the everyday movements of people who are suspected of no crime whatsoever.
That distinction, between targeted surveillance of a suspect and continuous collection of information about the general public, sits at the heart of the civil-liberties concerns surrounding these systems.
ShotSpotter Shows the Risks of Algorithmic Police Alerts
Oliver also examined acoustic gunshot detection systems, commonly associated with ShotSpotter. The technology uses microphones positioned throughout communities to identify sounds believed to be gunfire and rapidly alert police.
In theory, the system could help officers respond to shootings that otherwise might never be reported. In practice, however, distinguishing gunshots from fireworks, construction equipment, vehicles and other loud noises can be difficult. Records examined in Minneapolis covering thousands of ShotSpotter activations found that a large majority of police responses produced no documented evidence of a gun-related crime, while only a small percentage ultimately resulted in arrests.
That creates a problem more serious than a computer simply producing an incorrect result. When an algorithm flags a sound as gunfire, armed police officers may be dispatched into a neighborhood believing that a shooting has just occurred. The consequences of a false positive therefore extend far beyond bad data. Officers responding to what they believe is an active firearms incident may arrive with heightened expectations of danger, increasing the stakes for everyone they encounter.
Automated License Plate Readers Can Build Detailed Movement Histories
Automated license plate readers represent another major component of the expanding surveillance infrastructure. Systems operated by companies such as Flock Safety photograph passing vehicles and record information including license plate numbers, vehicle characteristics, locations and timestamps.
A single observation may seem relatively insignificant. A vehicle passed through an intersection at a particular time. But thousands or millions of observations collected across interconnected cameras can produce something far more revealing: a detailed history of movement.
That information can potentially show where someone lives, works, shops, worships, receives medical care or spends their evenings. It can reveal which neighborhoods a person regularly visits and, depending on the scale of the network, allow investigators to reconstruct substantial portions of someone’s movements.
The concern is therefore not simply that police can identify a stolen vehicle. It is that the same infrastructure capable of finding that stolen vehicle can also record the movements of enormous numbers of people who are not suspected of committing any crime.
False Reads Can Turn Into Dangerous Police Encounters
Accuracy creates another serious problem. Oliver revisited the disturbing 2020 traffic stop involving Brittney Gilliam and four children in Aurora, Colorado. Police believed Gilliam’s vehicle was stolen and conducted a high risk stop, ordering occupants out and handcuffing members of the family. The vehicle was not stolen. Authorities had confused information associated with a stolen motorcycle from another state.
The incident eventually resulted in a $1.9 million settlement. The case demonstrates why automated surveillance alerts cannot be treated as unquestionable evidence. A computer can make a mistake in milliseconds, but when that mistake is transmitted to armed officers, the consequences can become immediate and traumatic.
Flock Safety Has Become a Major Flashpoint
Flock Safety has become one of the central companies in the national debate over automated license plate readers. Law enforcement agencies argue that these systems can help locate stolen vehicles, identify suspects and find missing or endangered people. Civil liberties organizations counter that large networks of interconnected readers can create a de facto vehicle tracking system without requiring police to establish individualized suspicion before the information is collected.
Both realities can exist simultaneously. The technology can help solve legitimate crimes while also creating surveillance capabilities that would have been extraordinarily difficult and expensive for local police departments to build only a generation ago.
Stingrays Bring Surveillance Directly Into Americans’ Phones
Cell site simulators, commonly known as Stingrays, push the surveillance debate directly into Americans’ phones. These devices imitate legitimate cellular towers, causing nearby phones to interact with them. Police can use the technology to locate or identify a targeted device, but the equipment can also interact with phones belonging to innocent people in the surrounding area.
Oliver highlighted how law enforcement agencies have historically operated these systems under considerable secrecy, including through restrictive agreements and investigative practices that made it difficult for defendants and the public to determine when the technology had been used. That technology illustrates a fundamental transformation in surveillance.
Traditional investigations generally begin with a suspect and then seek legal authority to surveil that individual. Modern mass surveillance infrastructure can reverse that sequence. Governments can collect enormous amounts of information first, store it in searchable databases and determine later which person they want to investigate. Instead of suspect, warrant, surveillance, the system increasingly risks becoming surveillance, database, search, suspect.
Predictive Policing Can Reinforce Existing Biases
Predictive policing takes that transformation another step further. These systems analyze historical crime and police data in an attempt to identify geographic areas or individuals supposedly at elevated risk of future criminal activity. The concept sounds scientific, but algorithms inherit the characteristics of the data used to build them. If particular neighborhoods were historically policed more aggressively, those neighborhoods naturally produced more police reports, stops and arrests.
Feeding that information into a predictive system can cause the algorithm to identify the same communities as areas requiring additional policing. That can create a self reinforcing feedback loop. More police are sent into an area because historical police data indicates that the area generates significant police activity. Those additional officers conduct more stops and investigations, generating even more police data.
That new information is then returned to the algorithm, which appears to confirm its original prediction. The technology can consequently give old policing patterns a veneer of mathematical objectivity without eliminating the biases contained in the original data.
Florida Already Saw How Predictive Policing Can Go Wrong
Florida has already experienced one of the country’s most controversial examples of predictive policing. In Pasco County, the Sheriff’s Office developed an intelligence led policing program designed to identify people considered likely to commit future crimes. Investigative reporting later found that individuals placed within the system could receive repeated visits from deputies and scrutiny over relatively minor code violations and other issues.
Nearly 1,000 people reportedly entered the program over several years, including minors. The controversy became a powerful example of what can happen when an algorithmic assessment begins influencing how law enforcement treats people before they have committed a new crime.
Oliver’s Rialto Records Request Revealed Troubling Patterns
One of the most striking sections of Oliver’s investigation involved Rialto, California. Last Week Tonight filed public records requests seeking information about the police department’s use of Fusus and said it ultimately received a massive 30,413 page audit log. According to the show’s analysis, the records revealed officers accessing cameras inside local businesses, including fast food restaurants, as well as significant activity involving cameras at a Days Inn. Oliver specifically highlighted extensive access involving a hotel pool camera.
Those findings raise an important question that extends beyond Rialto: once police departments receive access to thousands of privately owned cameras, who is auditing what officers actually watch?
The Weakest Link May Be Oversight After the Technology Is Purchased
That question exposes one of the weakest points in America’s emerging surveillance infrastructure. The public debate frequently occurs when a city purchases technology, but far less attention is paid to how that technology is used months or years later.
Procurement rules, access logs, retention policies, data sharing agreements, disciplinary procedures and independent audits are far less exciting than a demonstration showing police instantly locating a stolen vehicle. Yet those safeguards ultimately determine whether a powerful investigative tool remains focused on legitimate law enforcement purposes or quietly evolves into routine monitoring.
This Is Not an Argument Against Police Technology
The debate should not be reduced to the simplistic argument that surveillance technology is inherently bad. A license plate reader that helps locate an abducted child can be extraordinarily valuable. A camera providing officers with real-time information during an active shooting can save lives. A gunshot detection system that identifies an otherwise unreported shooting could potentially get emergency medical assistance to a victim faster.
These are legitimate benefits, and ignoring them would weaken the case for meaningful reform. But technological usefulness does not eliminate the need for constitutional safeguards. In fact, the more powerful a surveillance system becomes, the stronger the oversight surrounding it should be.
Stronger Rules Are Needed Before These Systems Become Permanent Infrastructure
Governments should establish clear rules determining when surveillance databases can be searched, how long information can be retained, which agencies can access it, whether information can be shared across jurisdictions and what happens when officers misuse the system.
Automated alerts should require human verification before officers take potentially dangerous enforcement action, and independent audits should determine whether systems actually deliver the public safety benefits vendors promise. The public should also know when surveillance technologies are being purchased in the first place. Major surveillance contracts should not quietly move through local government without meaningful public debate simply because the technology is marketed as a crime-fighting upgrade.
Americans Can Investigate What Their Own Police Departments Are Using
Oliver pointed viewers toward the Electronic Frontier Foundation’s Atlas of Surveillance, a national database designed to help Americans identify surveillance technologies reportedly being used by law-enforcement agencies in their communities. The resource is particularly valuable because much of America’s surveillance infrastructure has developed locally rather than through one highly visible federal program. Residents may have no idea that their police department operates license plate readers, drones, camera networks or other monitoring technologies until journalists, watchdog organizations or concerned citizens begin examining municipal contracts and public records.
Public Records Can Reveal What Police Departments Do Not Advertise
For news organizations, those records can provide an extraordinary window into how modern policing actually operates. Contracts, invoices, procurement documents, audit logs, vendor communications, retention policies, user-access records and data-sharing agreements can reveal considerably more than a police department’s public description of a system. Oliver’s Rialto investigation demonstrates why those records matter. The public should not have to simply trust that extraordinarily powerful surveillance technology is being used responsibly when systems can be independently audited.
America Is Building Surveillance Infrastructure One Local Contract at a Time
The larger story is that America does not need to pass a single law creating a national surveillance state for something resembling one to emerge. It can be constructed incrementally through thousands of local decisions. One city installs license plate readers. Another builds a Real-Time Crime Center. Another connects private security cameras. Another purchases acoustic sensors. Another acquires cellphone tracking equipment. Another deploys predictive algorithms. Each purchase can be presented individually as a reasonable law enforcement upgrade. Connect enough of those technologies, however, and the result becomes fundamentally different from the individual components.
Surveillance Technology Is Becoming More Powerful Faster Than the Law
Cameras identify vehicles. License plate readers reconstruct travel. Cell site simulators locate phones. Private camera networks expand visual coverage. Algorithms search historical records. Artificial intelligence increasingly makes massive databases easier to analyze.
What once would have required teams of detectives physically following a person can increasingly be accomplished through a computer terminal. That is why the surveillance debate is ultimately larger than Flock Safety, Fusus, ShotSpotter, Stingrays or any individual police department. Cameras will become cheaper. Artificial intelligence will become more capable. Data storage will continue expanding. Surveillance networks will become increasingly interconnected. The technical ability of governments to monitor people is almost certain to increase.
Americans Should Not Have to Choose Between Privacy and Effective Policing
The unresolved question is whether democratic oversight will advance at the same speed. Americans should not have to choose between effective policing and basic privacy. Police departments can use modern technology while still operating under warrants, transparency requirements, strict retention limits, accuracy standards, public audits and meaningful consequences for abuse. Those protections are not obstacles to legitimate law enforcement. They are what separate targeted investigation from indiscriminate surveillance.
The Real Question Is Who Controls These Systems Tomorrow
The most important question surrounding these technologies is not what today’s police chief promises to do with them. Surveillance infrastructure can remain in place for decades, surviving elections, administrations, leadership changes and shifting political priorities. The more consequential question is what everyone who gains access to that infrastructure tomorrow will be capable of doing with it.
That is the warning underneath Oliver’s comedy. America is assembling an extraordinarily powerful surveillance network one camera, one database and one municipal contract at a time. The technology itself is moving quickly. The law needs to catch up.
