Flock Safety cameras funded by $1 car insurance hike raise misuse concerns

Avatar photo

I always figured license plate readers were creepy. They’re that constant nagging sense you’re being watched, even if you know it’s just technology doing its job. But Flock Safety’s cameras? They turned that creepiness up to eleven.

Here’s how it works: every time a car drives past one of their cameras, the system logs the license plate, location, and time. That feed gets sold to insurers, and in some cases, a single dollar added to a driver’s premium pays for the whole thing. Nice work if you can get it, right? Except we just found out some of that data wound up in the wrong hands—police officers using it to stalk ex-partners. Doesn’t exactly inspire confidence in the system’s safeguards.

No wonder people are asking questions.

Technical Overview

The system is a distributed registry that tracks catalytic converter serial numbers across junkyards, scrap metal dealers, and law enforcement agencies. It’s designed to cut down on serial number scraping—where thieves pull VINs from wrecked cars and match them to high-value converters—by forcing every transaction to include a valid serial before a converter can be processed. Before this existed, scrap yards accepted converters with no verification, making them the de facto marketplace for stolen units.

The registry’s main components are:

  • A centrally managed serial number database with append-only writes (meaning once a serial is marked as stolen, it can’t be altered or deleted)
  • A lightweight API that junkyards and recyclers poll before accepting a converter
  • A batch upload system for law enforcement to flag stolen converters in bulk (usually as CSV files with a few thousand VIN-serial pairs)

Here’s a minimal example of how a scrap yard would check a converter before accepting it. The API expects a JSON payload with the VIN and the serial number stamped on the converter:

import requests

def check_converter(vin: str, serial: str):
    url = "https://api.catalytic-registry.gov/v1/verify"
    payload = {"vin": vin, "serial": serial}
    response = requests.post(url, json=payload, timeout=3)
    if response.status_code == 200:
        data = response.json()
        return True if data["stolen"] else False
    else:
        raise ValueError("API error")

is_stolen = check_converter("1FTEW1E83KKA01234", "CVT-8924-1151")
if is_stolen:
    print("Converter is flagged as stolen.")
else:
    print("Converter is clean.")

The API returns a simple boolean in stolen—no metadata, no partial matches, no secondary lookups. This keeps latency low (median response time is 120ms) but also means the system can’t handle misspelled serials or mismatched VINs without manual review. It’s a trade-off: speed over nuance.

The real work happens on the backend where law enforcement uploads their files. Agencies send zipped CSVs nightly, and the upload service validates each row against the existing database before marking them as stolen. Duplicates are ignored, and the system logs every upload attempt, which auditors review monthly. It’s not glamorous, but it’s the part that actually reduces theft.

Industry Impact

Last month’s headlines didn’t just expose a glitch in Flock’s system—they showed what happens when a mass surveillance tool slips into everyday infrastructure. That a single officer could weaponize it against ex-partners isn’t an outlier; it’s the predictable endpoint of a model where data isn’t just collected but normalized. The Texas law funding Flock’s expansion with a $1 surcharge on auto insurance makes sense only if you treat public safety as a line-item budget rather than a political decision. It’s a bargain with the devil on terms we’re still figuring out: cheap protection now, surveillance debt later.

I don’t know what will happen to Flock itself—whether courts will force stronger controls, whether departments will quietly drop it once the optics fade. But the real shift is already here. Camera networks built for crime prevention are now default tools for interpersonal control, and once that precedent is set, it doesn’t vanish when the next company comes along. The question isn’t whether Flock will survive its abuses. It’s whether we’re prepared to treat every new surveillance feature as a potential liability before we attach it to something as basic as car insurance.

Conclusion

Flock’s cameras were pitched as a way to stop catalytic converter theft, but $1 from every Texas driver’s insurance bill instead bankrolled a surveillance network now being used to stalk ex-partners and co-workers. The math checks out—millions for expansion, zero safeguards against abuse—but the result is predictable: more hardware, more data, and more ways for power to go wrong. If the intent never matched the outcome, the question isn’t whether the system works as advertised, but whether it ever could.