Flock Safety says it does not do facial recognition. That is a clean, public commitment, and it sounds reassuring until you understand what it cannot actually prevent. This pattern of covert surveillance app infrastructure targeting civilians has surfaced in other contexts as well.
A sales email obtained through a public-records request by DeFlock Johnson City and reported by 404 Media exposes a specific gap. That gap sits between a company’s stated limits and the data those limits leave behind.
The Sales Pitch That Revealed the Gap
A May email to a Tennessee deputy police chief shows how third-party vendors can market capabilities that the original platform declines to offer.
In May, a VIDIZMO salesperson emailed Johnson City, Tennessee Deputy Police Chief Michael Adams pitching a platform called Intelligence Hub. According to 404 Media, the pitch described a system that could combine Flock Safety ALPR data, Axon body-camera footage, and other evidence into one searchable environment , reminiscent of how apps have been caught secretly tracking users well beyond their stated platform limits. From there, investigators could search by face, vehicle, or object. Johnson City police say they did not take the follow-up call.
VIDIZMO CEO Nadeem Khan confirmed to 404 Media that the specific Flock-to-VIDIZMO export tool described in the company’s marketing has not been built. Flock was not involved in, or informed about, the proposed integration, according to Khan.
This is a company marketing a prospective capability, not announcing a deployed one. That distinction matters, and it does not make the situation easier to dismiss.
The Problem Is the Architecture, Not Just the Feature
The real issue is not what VIDIZMO has built; it is what the surveillance data ecosystem makes possible for any vendor willing to try.
Here is the mechanism worth understanding. Flock’s privacy restrictions apply to what Flock itself builds into its devices and platform. Once an authorized police agency exports clips or plate reads under its own access rights, those files can enter a separate vendor’s system entirely. Khan described this to 404 Media as an agency-controlled import, characterizing VIDIZMO as a downstream processor rather than a Flock partner.
VIDIZMO’s broader platform documentation describes face matching against watchlists, behavioral analysis, object tracking, trespassing detection, and demographic classification. That last category includes estimated age, gender, and race across seven categories listed in the company’s materials: White, Black, Indian, East Asian, Southeast Asian, Middle Eastern, and Latino Hispanic, according to 404 Media.
Privacy expert Chris Gilliard, quoted by 404 Media, called the promotion of race, age, and gender filtering “appalling.” He added that race and gender are not static categories that computers can reliably determine.
Facial recognition systems have documented accuracy disparities across racial and gender groups. Building investigative tools around classifications the technology gets wrong most often for already-marginalized people is not a safeguard problem. It is a design problem.
VIDIZMO claims its system would require case numbers, log user actions for audits, and support face redaction for bystander privacy. Those are company claims, and they have not been independently verified.
Think of it like a streaming platform licensing content it did not produce: Flock builds the infrastructure, agencies collect the footage, and a third-party vendor arrives with tools the original company publicly declined to offer. The original provider keeps its clean reputation. The capability gets marketed anyway.
Gilliard argued, according to 404 Media, that once a mass-surveillance infrastructure exists, another company can layer more invasive capabilities on top of it even when the original provider refuses to offer them. San Jose has a Flock problem that illustrates exactly this community concern: when surveillance infrastructure becomes modular, responsibility for the most invasive capabilities moves from vendor to vendor. No single company ends up clearly accountable for the whole system.
Flock and Axon did not respond to 404 Media’s requests for comment. Whether any of this is legal in a given jurisdiction depends on applicable procurement contracts, data-use policies, and local laws. None of those have been independently verified for the scenarios described here.
The question worth asking your local police department is not whether Flock does facial recognition. It is what happens to the data after it leaves Flock’s platform.
