The Trump Card of Criminal Investigation Becomes a Stalker's Tool: What Japan Should Learn from the U.S. ALPR Issue

A small box-shaped camera standing by the roadside. It automatically reads the license plates of passing cars and converts the information such as date, time, location, and car model into searchable data.

Originally, this device was introduced for purposes such as discovering stolen cars, searching for missing persons, and tracking vehicles used in robberies or murders. However, in the United States, this "eye chasing criminals" has increasingly been revealed to be used by police officers themselves as a tool to stalk their partners or ex-lovers.

At the center of the issue is Flock Safety, which deploys automatic license plate reading cameras across the United States. The company's system installs inconspicuous cameras on roads and parking lots, making vehicle passage records searchable in the cloud. Each image is merely a snapshot. However, by linking the records of numerous cameras in chronological order, one can infer when a car leaves home, which workplace it heads to, and whose house, hospital, or religious facility it visits.

It's just reading numbers, not monitoring people—such explanations might hold as long as the data is scattered. But the moment dots become lines, and lines become maps of daily life, it becomes difficult to separate vehicle information from personal behavior history.

A police officer who searched for a partner hundreds of times

A symbolic case reported by Mother Jones occurred in Milwaukee, Wisconsin.

Police officer Josue Ayala reportedly searched the license plate of a person he was dating over 200 times in about two months in 2025. He even investigated the vehicle of that person's ex-partner. Ayala was prosecuted for abuse of authority and received probation after resigning.

However, another police officer, who was in a position to investigate this incident, was also suspected of using the same system to track individuals in his personal life. This illustrates a situation where those investigating abuse were themselves engaging in similar acts.

According to an investigation by the Institute for Justice into reports and public records, at least 24 cases nationwide were confirmed where police officers were suspected of using ALPR to track individuals they were romantically involved with. Most occurred after 2024, with many involved officers facing criminal prosecution, dismissal, or resignation. Importantly, only a few of these cases were initially discovered through internal audits, with most being revealed following victim reports.

This is not just a story of "a few unscrupulous police officers." It suggests that even if abnormal searches were repeated dozens or even hundreds of times, they might not have stopped unless the victim noticed the anomaly, gathered evidence, and reported it to the police.

The problem lies more with the "power to search" than with the cameras

The danger of ALPR is not just in the cameras capturing roads. More fundamentally, it lies in who can search the accumulated data.

To track a specific person using regular security cameras, a human must review numerous videos, which requires time and manpower. On the other hand, if license plates are databased, simply entering a number in the search field can instantly retrieve sighting records from multiple locations.

In other words, what used to require considerable effort can now be done in seconds from a desk or patrol car terminal.

Technology reduces the cost of criminal investigations. At the same time, it also reduces the cost of stalking and abuse of authority.

Moreover, those being tracked do not receive notifications that their car has been searched. Even if someone checks their number dozens of times, the person remains unaware. Although there is a system to input "investigation" or case classification as the reason for the search, if no one verifies whether the explanation is true, it is practically close to self-reporting.

The gap between a system designed to trust those with authority and the reality that those with authority can err was exposed in this incident.

Flock argues "abuse can be recorded"

Flock Safety acknowledges the possibility of abuse. However, they explain that search histories are recorded for each user, allowing for later confirmation of who searched for what, when, and for what reason. They have also established a permanent audit log that cannot be deleted and a feature called Audit Assistance to detect suspicious usage patterns.

The company claims that the majority of searches by over 140,000 monthly users are legitimate investigations, and that the reported abuses are rare in the overall context. They also emphasize that it was because audit logs existed that they could prove misconduct and discipline police officers in some cases.

There is a certain rationality to this rebuttal. Systems that leave operational histories can sometimes make it easier to prove misconduct than using paper documents or orally obtained information.

However, "leaving evidence" and "preventing harm early" are separate issues.

Even if logs are found after a police officer has searched for an ex-partner 200 times and is disciplined, the fear and danger the victim felt before that do not disappear. Audit logs are useful for post-fire cause investigations. But that alone does not make them fire alarms.

What is truly needed is a system that immediately detects unnatural repeated searches for the same number, inquiries about individuals unrelated to the case in charge, and access outside working hours, and confirms them by an independent supervisory department.

Reactions on social media: "It was indeed abused"

 

Reactions to this issue on English-speaking social media and forums are largely divided into three.

The first is a strong distrust that "the explanation to trust internal safety measures is meaningless."

In Reddit's tech community, there were numerous comments pointing out that merely requiring a reason to be entered at the time of search allows users to write any explanation they like. Questions arose about whether so-called "internal controls" are actually based on the conscience of police officers, who regularly checks abnormal search counts, and who acts when they are discovered—the police chief, internal affairs, or an external agency.

The second opinion is that the system itself should be dismantled.

From this standpoint, the problem is not the lack of audits but the fact that the movement records of citizens without any suspicion of crime are being massively stored from the start. The reaction is that instead of "increasing supervision," the very system of constant tracking in a free society should be eliminated.

The third is a realistic opinion that acknowledges the effectiveness of the system in criminal investigations while advocating for strengthened audits.

There are situations where ALPR is effective in discovering stolen cars or vehicles involved in major incidents. Since the recent misconduct was ultimately proven through logs, the argument is to not dismantle the system entirely but to expedite the detection and handling of misconduct. In Reddit discussions about the Wisconsin incident, opinions like "it should be abolished because it's constant surveillance" and "since it helps solve actual crimes, a design to quickly detect abuse is necessary" are both present.

These are not public opinion surveys but reactions on social media where participants voluntarily post. Nonetheless, what is noteworthy is that beyond the traditional dichotomy of "security or privacy," specific discussions about "who audits the audit logs" are spreading.

More municipalities are reconsidering contracts

The backlash against Flock is not limited to criticism on social media.

According to the ACLU, over 55 municipalities have terminated contracts with Flock in the past year, driven by concerns over immigration investigations and tracking related to cross-state abortions. The ACLU points out that the network conducts over 20 billion vehicle scans per month, and the key regulatory concerns are data retention periods, sharing partners, and usage purposes.

What emerges here is the risk that "a camera installed by one city" does not remain just that city's issue.

Even if local police introduce it to search for stolen cars, if agencies from other regions or states can search it, it might be used for purposes residents did not anticipate at the time of installation. If an action legal in one area is punishable in another, nationwide data sharing can easily cross legal boundaries.

Local security equipment grows into a de facto wide-area tracking network. This is another core issue of the Flock problem.

It's not just a "distant country's story" for Japan

The systems in the U.S. and Japan's police systems are not the same. It is not possible to simply equate Flock's private company-led network with the system operated by Japanese police.

Even so, it is difficult for Japan to treat this issue as someone else's problem.

In Japan, an "Automatic Vehicle Number Reading System," commonly known as the N-System, has been established to automatically read the numbers of passing vehicles and match them with wanted vehicles. The National Police Agency explains that it is effective for early arrest of suspects and discovery of wanted vehicles because cars are used in crimes or escapes.

There are numerous security cameras at stations, shopping streets, apartments, stores, and roads. It is no longer uncommon for footage from private dashcams or store cameras to be used in investigations. In the future, as AI-based vehicle identification, searches for characteristics of people and clothing, and cross-referencing between multiple cameras become more advanced, individual footage will rapidly transform into "searchable behavior history."

In Japan, there is also a strong perception that "if you haven't done anything wrong, there's nothing to worry about." However, the people tracked this time were not criminal suspects. They were investigated by a state agency's investigation system merely because they dated or had dated a police officer.

The safety of surveillance technology is tested not when used by good managers, but when accessed by individuals with obsession, prejudice, or political intent.

Japan's personal information administration also recognizes the "risk of tracking"

Japan's Personal Information Protection Commission, in its report on cameras with facial recognition capabilities, lists concerns such as long-term tracking of specific individuals, automatic, indiscriminate, and massive information acquisition, difficulty for individuals to predict usage purposes, and the potential to cause behavioral inhibition.

This report focuses on facial recognition and is not a document directly regulating license plate reading systems. Nonetheless, the problem structure of being able to extensively track individuals' actions by combining data is common.

The commission also points out that it is necessary to not only comply with the Personal Information Protection Law but also to fully consider not infringing on portrait rights and privacy. In other words, even information that can be legally obtained can potentially infringe rights depending on the scope, purpose, and method of use.

Regarding license plate information, the idea that "because it's visible on public roads, it's free to collect" is insufficient. Seeing a car once and searching its passage records at multiple locations over weeks have completely different impacts on an individual's life.

Six conditions Japan should decide in advance

If Japan is to expand number recognition and AI cameras in the future, it will be too late to take measures after problems arise post-introduction.

First, searches should require specific case numbers or legal grounds, and not be searchable for vague reasons like "investigation" or "confirmation."

Second, there should be a system that automatically detects repeated searches for the same vehicle or access to cases outside one's responsibility, notifying an independent supervisory department rather than the police station the individual belongs to.

Third, it should not end with self-audits within the police station; a third-party organization should regularly verify the number of searches, purposes, violations, and disciplinary results. Statistics can be disclosed to citizens in a non-personally identifiable form.

Fourth, data unrelated to the case must be deleted as soon as possible. Leaving movement records of all citizens for a long time under the pretext that "it might be useful someday" increases the risk of leaks and misuse.

Fifth, data sharing across municipalities and prefectures should require individual approval and recording. The state of being shared nationwide just by the system's initial settings should be avoided.

Sixth, private use by staff should not be treated merely as a violation of service discipline but should be strictly investigated as stalking or abuse of authority. Including notification to victims, consultation support, and safety assurance in the system is also important.

Even if there are technical logs, they are meaningless without a responsible person to verify them. Even if there are rules, if violations are tacitly approved within the organization, they do not serve as a deterrent. And even if there is a disciplinary system, if victims have to discover misconduct on their own, it cannot be considered sufficient protection.

Security and privacy are not trade-offs

In discussions surrounding the Flock issue, there is often an argument that "to solve crimes, some level of surveillance should be accepted."

However, the real question is not a binary choice between criminal investigation and privacy.

It is about using technology effective for investigations while ensuring that unrelated citizens cannot be tracked. It is about being able to search quickly in emergencies, while private searches are quickly detected. It is about not only the implementing municipalities but also citizens themselves being able to verify the operational status.

What is required is not to discard technology or trust it entirely, but to design a system that controls power.

Cameras may not betray people. But the people behind the cameras can be jealous, obsessive, make mistakes, and sometimes abuse their authority.

That is why, when introducing surveillance systems, the premise should not be "only trustworthy people will use it," but rather "someone will inevitably attempt abuse."

In an era where cameras on roads continue to increase, what citizens need to know is not just the number of installations.

Who can search their information? Who monitors the fact that it was searched? When mistakes or abuses occur, will the person be informed?

It is society that truly needs to monitor the "eyes of the police" installed for safety.

---

Source URL

Mother Jones: Introduces the tracking of partners by Milwaukee police officers and the private use of Flock cameras across the U.S.
https://www.motherjones.com/politics/2026/07/police-losing-jobs-flock-cameras-alprs/

Institute for Justice: Compiled cases suspected of using ALPR to track romantic interests. Confirmed at least 24 cases, with many uncovered through victim reports rather than internal audits.
https://ij.org/police-have-reportedly-used-license-plate-readers-to-stalk-romantic-interests-at-least-14-times-in-recent-years/

404 Media: Detailed reports on multiple incidents where police officers repeatedly searched for ex-partners and their families using F