Written by Scott
Every few months, a new municipal battle erupts over automated surveillance. Right now, the lightning rod is Flock Safety’s automated license plate reader (ALPR) network.
On one side, civil liberties advocates warn that a nationwide, AI-powered camera mesh is eroding the Fourth Amendment and turning public movement into a permanent, searchable dragnet. On the other side, local police departments and neighborhood associations insist that these cameras are indispensable tools for catching carjackers, recovering stolen vehicles, and finding missing children.
I watch this debate unfold with a growing sense of frustration. Both sides have locked themselves into an unproductive, binary argument:
Either ban the technology entirely, OR
Accept unchecked mass data ingestion with minimal oversight.
We don’t have to choose between an unregulated surveillance state and 1980s-era analog policing. There is a pragmatic third path: embrace the objective, crime-solving power of high-resolution data, but demand the enterprise-grade compliance and audit architecture required to govern it.
The Baseline Fallacy: The Old Way Was Broken
Much of the pushback against modern surveillance suffers from status quo bias—the tendency to compare the risks of new technology against an idealized, flawless past that never actually existed.
Critics often talk about pre-digital policing as if it were inherently fairer and less intrusive. In reality, legacy policing was low-resolution, deeply subjective, and rife with catastrophic clerical errors.
I know this firsthand. Before modern digital verification systems existed, I was detained by law enforcement on three separate occasions:
The “Vague Description” Dragnet: In college, police stopped and detained about seven of us simply because we were all roughly 6’5” with brown hair—matching a vague physical description of a suspect. We sat on the curb until someone could physically come look at us.
The Clerical Identity Collision: In 1980, while applying for a vendor license in Ohio, a legacy database match erroneously flagged me as an armed robbery suspect. Detectives arrived, guns were drawn, I was handcuffed, and I spent six hours detained before the clerical mix-up was untangled.
The Absurd Administrative Lag: I was once detained by federal authorities investigating the murder of a prison guard and an escape that quickly resolved when someone finally checked basic records and realized I was five years old when the crime occurred.
Analog policing was already a dragnet; it was just a slow, imprecise, and biased one.
High-resolution digital data, exact timestamps, precise license plate matches, make, model, and visual verification, provides objective ground truth. It doesn’t just catch the guilty; it rapidly exonerates the innocent and prevents people from being held at gunpoint over vague descriptions or clerical typos.
Why the Public Is Still Uneasy
If high-fidelity data is vastly superior to legacy guesswork, why are communities actively protesting Flock?
I have cameras on my house at a busy downtown intersection. My neighbors do too. When an accident happens, a car is vandalized, or someone needs footage to defend against a bad traffic citation, we pull the footage and help out. Nobody protests my house.
The public pushback against enterprise ALPR grids isn’t about cameras existing in public spaces. It comes down to scale and topology:
A home camera is an isolated node. It captures an event at a single coordinate, and human friction is required to pull the footage for a specific, known incident.
A centralized ALPR mesh is a relational database. When you link tens of thousands of cameras across municipal, county, and state borders, you move from incident investigation to behavioral pattern-of-life mapping without a rationale purpose.
When you combine that scale with the state’s monopoly on coercive force, the public’s apprehension is rational. If an ordinary citizen can be tracked everywhere they go, who is auditing the people holding the search engine?
Manage the Risk, Enjoy the Benefits
The good news is that we do not need to invent new concepts to fix this. In the enterprise risk world, we have spent decades designing frameworks to manage highly sensitive, aggregated data storage locations.
If we want the safety benefits of high-resolution surveillance without the risk of an unaccountable surveillance state, we can and should mandate four baseline controls:
1. Cryptographic, Non-Repudiable Auditing
Standard vendor managed access logs are not enough. System access should be tied to append-only, tamper evident logs. Every query—who searched, what plate, when, and tied to what specific active case number—must be permanently recorded and structurally impossible to alter.
2. The Right to an “Accounting of Disclosures”
In heavily regulated industries, individuals have the right to request an audit log of who accessed their sensitive records. Why shouldn’t a law-abiding citizen have the right to request an accounting of when and why their license plate was queried in a public database?
Knowing that a search will eventually be visible to the subject is the single most effective deterrent against internal snooping, pretextual searches, and unauthorized snooping on ex-spouses or political rivals.
3. Independent Oversight with Real Teeth
Self-policing does not work. Internal affairs reviews of ALPR logs have historically been treated as administrative checkboxes. Oversight should be handled by independent civilian boards or state auditors with the power to revoke system access, levy fines, and refer bad faith queries for criminal prosecution.
4. Mitigating Automation Bias
An optical character recognition (OCR) match or out-of-state plate is an investigative lead, not absolute ground truth. Agencies must implement strict operational protocols requiring secondary human verification before patrol officers initiate a high risk felony stop. You don’t pull guns on an innocent driver because an algorithm misread an 8 as a B.
The debate over Flock cameras has been dominated by two extremes:
those who want to pretend modern technology has no place in public safety, VS
those who want to deploy high powered tracking tools without building the governance to match.
We can have safe, accountable communities without surrendering constitutional boundaries. The path forward isn’t tearing down the cameras and returning to the flawed, dangerous guesswork of the 1980s.
The path forward is demanding that the institutions using these powerful tools govern them with the transparency, rigor, and accountability that modern risk management demands.
The Fiscal Realities
One of the main reasons ALPR networks have exploded so quickly is their artificially low entry cost.
By stripping away the overhead of third-party auditing, independent oversight, and strict compliance infrastructure, vendors can sell turnkey hardware-as-a-service(HAAS) subscriptions for roughly $3,000 per camera per year.
Critics will immediately object that implementing the rigorous governance I am proposing, tamper-proof cryptographic logging, subject access disclosure systems, independent civilian audit boards, and strict training protocols, will significantly increase the per-unit cost.
They are correct. It will cost more. And it doesn’t matter.
From a risk and capital allocation perspective, this added cost is not a flaw; it is a vital filtering mechanism.
1. High-Incident Areas Easily Justify the Fully Loaded Cost
In high-crime corridors, commercial transit arteries, and persistent auto-theft hotspots, the return on investment for high-resolution visual data is massive. The direct financial recovery of stolen vehicles, the reduction in severe insurance claims, the mitigation of violent crime, and the savings in billable detective overtime easily absorb the cost of both the camera and the compliance apparatus required to govern it.
2. The Litmus Test for Camera Placement
Accounting for the “fully loaded cost” (Hardware + Software + Independent Governance) creates a natural, disciplined threshold for municipal deployment.
The rule should be simple: If an area does not have the incident volume or public safety risk to justify the cost of monitoring it responsibly, we have no business putting a camera there in the first place.
Right now, artificially cheap, un-audited cameras encourage indiscriminate municipal sprawl—installing lenses on quiet residential side streets just because the budget allows it. Forcing cities to budget for full oversight ensures cameras are placed only where real, measurable risk exists.
3. The Commoditization Curve
Starting with targeted, fully governed deployments in high-impact zones also aligns with standard technology lifecycles.
As edge-compute hardware, optical sensors, and standardized compliance software commoditize over time, the per-unit cost of both the cameras and the auditing tools will steadily drop. As the fully loaded cost declines, cities will be able to responsibly expand coverage to broader areas without ballooning public budgets or compromising on civil liberties.
The Fiscal Cautionary Note:
The fastest way to destroy public support for this technology is to turn it into a revenue-generation tool.
The moment ALPR is repurposed for automated ticketing, registration lapse fines, or municipal debt collection, it ceases to be a violent-crime deterrent in the eyes of the public and becomes an automated tax. The fiscal justification must stay anchored strictly to public safety, asset recovery, and investigative efficiency.
