On December 13, 2025, a masked gunman opened fire in a Brown University classroom, killing two and wounding nine. The shooter, who carried out the attack in a camera-free area, managed to escape. Days later, police got a lead on a gray Nissan with Florida plates. They ran it through Providence’s Flock Safety automated license-plate-reader network (ALPR), traced the car to a rental agency, and identified the shooter within hours. Providence officials called Flock “instrumental” to the investigation.

This, among other success stories, is largely the product of what Flock has done to a decades-old policing technology. It made ALPR cameras cheap and scalable, layered in AI-powered search (i.e “blue hoodie,” “gray Nissan”) and connected them across jurisdictions through a nationwide network. Its broader platform now includes video cameras, gunshot sensors, and drones.

Nine months later, Americans are now ripping Flocks’s ALPR cameras to the ground en masse. Cities are canceling contracts, Republicans and Democrats are pushing bans at both the local and national level, and “deflocking” has become enough of a movement that one viral post called it “the closest Americans have gotten to the Boston Tea Party since 1773.”

The backlash stems largely from some pretty ugly stories about how Flock’s system is being used.

During probably the worst two-week stretch of press in Flock’s history, founder Garrett Langley gave Pirate Wires his side of the story — why the technology works, where the abuses come from, and whether America can have Flock without becoming a surveillance state.

Creepy, Life-Saving Cameras

We started with the obvious question: does this stuff actually work?

According to Langley, yes. “In July alone, [Flock] helped detect roughly 1,100 missing people, 6,000 felony warrants, and more than 20,000 stolen vehicles” nationwide.

“There’s really no debating the efficacy of the technology,” he told me

A primary reason for that success — and, simultaneously, people’s fears about the tech — is Flock’s enormous scale. Its network spans more than 120,000 cameras across 6,000 communities, generating roughly 20 billion plate reads a month.

That’s a lot of flocking data.