Mass Surveillance, Police Misuse, and Who Controls Your Flock Cameras with Flock CEO, Garret Langley

Why does it matter? Because the platform tracking your license plate is also firing the cops who abuse it.

Flock Safety expected its new abuse-detection tool to surface a handful of bad officers. Instead, it found enough to make the feature mandatory across all 6,000 city customers. CEO Garrett Langley is now publicly working through what mass surveillance infrastructure looks like when the vendor decides to police itself — and the specific numbers that come out of that conversation should change how the debate is framed.

• Flock's AI audit tool found so many dirty cops it shifted from opt-in to mandatory — with chiefs in Georgia publicly announcing mass firings • 90% of crimes solve within 7 days of data retention; the remaining 10% are homicides and unreported rapes • Flock bans facial recognition, video capture, and people searches — most competitors impose none of these limits • State retention laws span 3 minutes (New Hampshire) to 5 years (New Jersey), with federal regulation near zero

Flock's AI audit tool went looking for dirty cops — and found far more than anyone expected

Four months ago, the tool shipped as an optional feature. It's now mandatory across every customer. That shift tells you everything about what turned up when they started looking.

Langley explains the pattern: Flock solved over a million crimes last year, which means the company has a precise model of what normal usage looks like. Abuse has a different signature. An officer stalking someone searches the same plate repeatedly — and never puts it on the hot list, never loops in other shifts, because they don't want anyone else to know. A cop working a real case does the opposite. The anomaly is easy to flag once you know what you're looking for.

What the tool surfaced when it went live: "a ton." Chiefs in Georgia ran with it. One publicly announced he'd fired nine officers — the harshest disciplinary outcome under state law — specifically for abusing the system.

Langley's position hardened: "If a customer doesn't want to be held accountable, I don't think I want them as a customer."

The primary objection to surveillance infrastructure has always been that it creates a honeypot for bad actors. Flock's answer is to build the honeypot deliberately — and make the vendor, not just the department, responsible for catching who reaches in.

90% of crimes solve in 7 days — but the 10% left over are homicides and rape cases

Ninety percent is a reassuring number. Then you ask what's in the other ten.

The homicide where no 911 call ever came in because the body took five days to find on a welfare check. The rape victim who felt guilt, who felt shame, and didn't call until a week later because she wasn't ready. "The 10% we're talking about," Langley says, "is the 10% you read about in the news." These aren't edge cases — they're precisely the crimes that demand longer retention, not shorter.

Flock recently shifted its default from 30 days to 7, and Langley is clear this is a real tradeoff. State law complicates it further: New Jersey mandates five years; California and Washington have set maximum caps; New Hampshire deletes at three minutes. Most communities land somewhere in between, but where exactly is a values question, not a technical one.

A retention vote is a direct answer to one specific question: which crimes are we comfortable leaving unsolved? Communities making that decision on bad information — or no information — are making it wrong.

The company absorbing all the surveillance backlash is the one that banned facial recognition

No facial recognition. No video. No looking inside cars. No people searches by name. Flock built these restrictions directly into the product — in direct conflict, Langley notes, with its biggest competitors, who do all of it.

One of Flock's first cities to walk away did so over facial recognition fears. They googled "ALPR facial recognition," landed on a competitor's marketing page, and canceled Flock. "That's on me," Langley says. The outcome: the industry's most restrained player absorbs the public heat while vendors with no comparable limits operate below scrutiny. Langley's bet is that it eventually catches up to them. Right now, it hasn't.

The opposition also skews predictably toward the affluent. People with gated entries, security guards, alarm systems, and doorbell cameras can treat the privacy tradeoff as abstract — they've already solved the safety problem privately. An Illinois legislator who was one of Flock's most aggressive critics changed her position publicly after police used a Flock camera to catch the person who shot at her house.

Where you stand on public surveillance cameras often reflects what private security you've already paid for.

Drones-as-first-responder are already deployed at scale — and the legal framework is nearly nonexistent

A woman calls 911 from a gas station: man walking back and forth with what looks like a gun. Old model — one officer races over, heart rate spiking, weapon drawn, someone might die. New model: a drone launches from a rooftop dock, flies 60 miles per hour at 400 feet, reaches the scene in under a minute, zooms in at 40x magnification. It's a lighter. The dispatcher calls the woman back. She's okay.

Langley calls drones "probably our fastest growing business unit" — and then acknowledges that state regulation on drone-as-first-responder is nearly nonexistent, and federal oversight covers only air safety. Who can authorize a launch, what footage is retained, how long it lives, who can review it — none of this has a settled legal framework. Flock sets the defaults. Cities rarely revisit them.

The cameras debate has years of litigation, protest, and case law behind it. The drone debate has barely begun.

Flock is the de facto policymaker for 6,000 jurisdictions — and Langley knows it

The accountability chain is short: elected officials defer to police chiefs, who defer to Flock. Langley doesn't obscure this. "The responsibility falls on us to give those strong defaults and then allow them to pivot against it." Elected officials can't be expected to understand surveillance architecture any more than water treatment systems — they rely on the vendor to educate them on what is best.

That makes Flock's product decisions functional law for billions of data points about people's movement across 6,000 cities. It's also why making the audit tool mandatory mattered: contractual provisions allow deviation; defaults enforce it. Most cities will never touch either.

Scrutinize a surveillance vendor's defaults as aggressively as their contractual terms. Most jurisdictions won't — and most vendors are counting on that.

Surveillance infrastructure that polices itself is the exception — not yet the model

Flock didn't set out to build a police accountability institution. It built a crime-solving tool that accumulated enough behavioral data to recognize anomalies — and then pointed that pattern recognition at its own customers. That loop works right now because the company is choosing to apply it. The harder question, as the platform expands into drones, AI analysis, and real-time behavioral assessment, is whether a vendor's good intentions can substitute for a governance framework that doesn't yet exist. One company setting defaults across 6,000 cities isn't oversight. It's a gap. The next fight isn't over cameras.

---

Topics: surveillance, public safety, privacy, law enforcement technology, license plate readers, drones, police accountability, AI in policing, civil liberties, govtech, smart cities, startup