‘No Kings’ Movement Taking Its Fight To Flock Cameras

Military analysts monitor data on multiple screens in a command center — Photo: Gorodenkoff / Shutterstock

What makes automatic license plate reader systems consequential is not the camera bolted to a pole; it is the network effect that turns millions of fleeting roadside sightings into a searchable ledger of movement that any connected agency can query—and, in practice, often share.

At a Glance

Flock Safety markets a nationwide ALPR network that captures passing vehicles and stores location-linked data accessible to law enforcement.
Independent reporting and public records show agencies using Flock to run protest-related searches and to assist immigration enforcement, commonly via local-agency intermediaries rather than direct ICE logins.
Flock denies direct ICE access and frames access control as a community decision; the strongest evidence concerns indirect pathways, not a standing ICE contract.
The civil-liberties risk stems from scale, retention, and cross-jurisdiction sharing—distributed infrastructure that functionally enables mass location tracking.

What Flock’s network actually is

By design, ALPR systems turn public roads into data sources. Flock Safety’s own materials and allied civil-rights statements describe the product not as a single agency’s tool but as a national network: cameras capture images of passing cars, extract plate numbers, and persist time-and-place-linked records in a database available to law-enforcement entities. That geometry—widespread capture, centralized storage, and federated access—is what transforms episodic detection into a historical trail. On its public site, the company emphasizes that it does not provide direct access to U.S. Immigration and Customs Enforcement (ICE), but acknowledges that access decisions ride with customer agencies; if a community permits federal partners to obtain results through them, the system accommodates that pathway.

This distinction—direct versus indirect access—matters legally and politically, but it does not change the system’s technical affordances. If a local police department can search the network and relay hits, the operational effect for the target is similar to an ICE analyst typing the query. Whether the keystrokes originate in a federal cubicle or a local crime center, the network’s value proposition is the same: find the plate across place and time.

How indirect access works in practice

Public-records-driven reporting documents that federal immigration enforcement has in fact benefited from such indirect pathways. Summaries of records indicate ICE officials asked state and local departments to run Flock searches on their behalf; one compilation pegs the number of immigration-related lookups in the thousands, with specific instances in Florida where wildlife officers and other state entities conducted Flock queries for ICE purposes. Separately, reporting on a letter attributed to Sen. Ron Wyden describes access by multiple federal agencies, including ICE’s investigative arm, to the nationwide network, and cites search counts for Homeland Security Investigations—granular figures that move this discussion beyond abstraction.

Flock maintains that federal access occurs only if a customer agency permits it. That caveat is consistent with the records: the contested behavior is not a secret API for ICE, but the routinized practice of federal agents tasking local partners to run the query and share the output. In short, governance by memorandum replaces governance by firewall. Communities that deploy ALPR infrastructure therefore need to understand that their sharing rules—not vendor branding—determine whether federal enforcement rides along.

Protest monitoring and the First Amendment problem

ALPRs do not read political intent; they read plates near a place and time. That agnosticism is the privacy risk. Analyses by digital-rights investigators who obtained large volumes of search logs show that dozens of agencies ran queries linked to protest activity over a ten-month span, including searches pegged to “No Kings” demonstrations; they attribute hundreds of protest-related searches across more than 50 federal, state, and local entities, with nineteen agencies conducting dozens tied specifically to those protests. Local reporting corroborates this pattern at the department level—Tulsa police, for example, logged at least 38 Flock searches related to 2025 protest events, including the June “No Kings” actions.

Defenders counter that such instances represent a small fraction of total customers, emphasizing that the absolute number of agencies explicitly referencing protests is tiny relative to Flock’s footprint. That percentage argument misses the structural issue: when a network makes retrospective movement queries trivially easy, edge-case uses propagate. Constitutional exposure does not hinge on market share; it turns on whether government can, without a warrant or narrow tailoring, assemble location histories around expressive activity. Several courts have begun to wrestle with how camera density, retention periods, and cross-agency search capabilities bear on the Fourth Amendment analysis; as the network densifies, the jurisprudence tends to sharpen its scrutiny.

Direct versus indirect ICE access: what the evidence does and does not show

Precision matters. The record surfaced here does not establish a standing Flock–ICE contract or an ICE-operated console into Flock’s systems. Flock explicitly denies direct access, and nothing in the cited materials cleanly refutes that claim. What the reporting and public records do document is recurrent, operationally meaningful use of the network for immigration enforcement via local and state partners—thousands of lookups in some jurisdictions, and specific tasking of agencies to run searches that ICE could not run directly. Federal access described in summaries of congressional correspondence, and state-level reporting on agencies assisting ICE, reinforce the same point: indirect access has been sufficient to achieve enforcement objectives.

The practical upshot for policy and civil rights is straightforward. If indirect pathways deliver the same enforcement outcomes as direct logins, communities cannot comfort themselves with vendor denials alone. They must interrogate interagency agreements, audit trails, and approval workflows; otherwise, “no direct access” is a semantic shield for a shared-operational reality.

The network effect: retention, sharing, and risk

Three design choices amplify risk in ALPR ecosystems. First, retention windows: even relatively short periods can be long enough to reconstruct attendance at a rally, a clinic, or a mosque. Second, cross-jurisdiction sharing: a plate seen in one town becomes a breadcrumb available to many, including distant agencies with different priorities. Third, query semantics: the ability to run geofenced, time-bounded searches turns broad dragnet capture into precise dossiers. Together, those levers convert anodyne “public-facing” photography into a de facto location-history service. Civil-rights coalitions have warned about this convergence for years, pointing specifically to immigration risks and due-process concerns when networked ALPR data flows into federal investigations.

What rigorous oversight looks like

Communities that choose to deploy ALPRs can narrow the aperture. The controls that matter are knowable and auditable: prohibitions on protest- or immigration-related searches; explicit bans on running queries for or on behalf of federal immigration enforcement; formal approval chains with named supervisors; immutable, regularly reviewed audit logs; tight retention with automatic deletion; and bright-line sharing restrictions that require public votes to alter. Equally important is external verification—public-records disclosures of aggregate query counts and reasons, independent technical audits of search capabilities, and periodic civil-liberties impact assessments. Without these, “policy” is a press release, not a control system.

Why this debate will not fade

ALPRs are part of a broader shift toward distributed surveillance, where ordinary infrastructure—streetlights, traffic sensors, cloud platforms—accumulates enough coverage and memory to approximate continuous tracking. In that context, the argument over whether 0.5% or 5% of agencies ran protest-related searches is a sideshow. The central question is whether a national network that can be queried across jurisdictions should be used to map political assembly or assist civil immigration enforcement through proxy channels. The evidence to date shows that, absent explicit guardrails, it will be. The technology’s trajectory is set by governance, not slogans. Communities that fail to decide are, functionally, deciding.