Flock Safety's nationwide network of automated license plate readers has transformed law enforcement capabilities across America, yet recent audit logs expose a darker reality: over 80 agencies systematically misused the system between June 2024 and October 2025 to conduct hundreds of discriminatory searches targeting Romani communities, abortion seekers, and political protesters without criminal justification. EFF investigations reveal how this surveillance infrastructure—spanning 83,000 cameras across 5,000+ communities—has become a tool for chilling constitutional rights, with minimal oversight enabling everything from racial profiling using slurs like "g*psy" to unauthorized Border Patrol access to local networks.

Key Takeaways

• Over 80 law enforcement agencies conducted hundreds of discriminatory searches using terms like "roma" and "g*psy" to racially profile Romani and Traveller communities between June 2024 and October 2025
• Agencies queried Flock's network of 83,000 cameras for women who "had an abortion," enabling warrantless tracking of reproductive healthcare seekers across state lines post-Roe
• Border Patrol gained unauthorized "back door" access to ten agencies' data between May-August 2025, while eight Washington state agencies enabled direct 1:1 sharing for migrant tracking
• Despite Flock claiming 700,000 crimes solved annually, minimal oversight and routinely ignored bias policies have enabled systematic abuse for political surveillance and First Amendment violations
• Sen. Ron Wyden and advocates demand warrant requirements, contract terminations, and opt-outs as Flock expands into AI video and voice surveillance technologies

Discriminatory Searches Against Romani Communities

Between June 2024 and October 2025, over 80 U.S. law enforcement agencies conducted hundreds of searches using explicitly discriminatory terms targeting Romani and Traveller communities. These searches included terms like "roma," "g*psy," "roma traveler," "possible g*psy," "g*psy ruse," "g*psy vehicle," and "g*psy group"—often without any connection to actual criminal investigations. Grand Prairie Police Department in Texas alone searched "g*psy" six times using Flock's "Convoy" feature specifically designed to track traveling groups.

The scale of this racial profiling extends far beyond isolated incidents. EFF analysis revealed more than 400 additional searches targeting Traveller communities, including Irish Travellers. Specific examples from audit logs include:

• Palos Heights PD: "g*psy" searched 24 times
• Multiple agencies using "roma traveler" without case numbers
• Grand Prairie PD: six "g*psy" Convoy searches to track movement patterns
• Widespread use of "g*psy ruse" and "g*psy vehicle" as search justifications

This digital discrimination reflects centuries of anti-Roma racism, from Holocaust genocide to modern bias. The U.S. State Department formally recognizes "Anti-Roma Racism" as a distinct form of persecution, yet law enforcement agencies have essentially industrialized this bias through Flock's searchable nationwide network. Historical parallels include NABI's infamous "G*psy crime" database, which similarly codified ethnic profiling into law enforcement infrastructure.

A 2020 Harvard survey found that 4 out of 10 Romani Americans report experiencing racial profiling by police. Flock's technology amplifies this pattern by bypassing local oversight policies and enabling instant nationwide searches based purely on ethnic identifiers. Oak Park, Illinois data reveals the broader impact: 84% of Flock-related traffic stops involved Black drivers, despite Black residents comprising only 19% of the population.

Agency responses to EFF's findings condemned the discriminatory language but exposed significant enforcement gaps. Palos Heights officials stated the searches "do not reflect our values," yet audit logs show no disciplinary actions or system changes were implemented. This mirrors broader patterns documented across law enforcement, where codes like "furtive movements" disproportionately target Black, Latine, and Native communities.

Tracking Abortion Seekers and Reproductive Rights Violations

Law enforcement agencies have weaponized Flock's network of 83,000 cameras to track women seeking reproductive healthcare. EFF's 2025 investigations uncovered queries specifically searching for individuals who "had an abortion," creating a chilling surveillance apparatus in the post-Roe landscape. These searches enable tracking of movement patterns across state lines without warrants or judicial oversight.

Flock's query system theoretically requires officers to input reasons and case numbers for nationwide searches. However, audit logs reveal these fields are routinely omitted or filled with vague justifications. A single abortion-related query can instantly scan thousands of cameras across multiple states, creating comprehensive movement profiles of healthcare seekers.

While Flock claims it doesn't deploy facial recognition technology, vehicle data provides nearly equivalent tracking capabilities. License plate captures include timestamps, vehicle make/model/color, and distinguishing features—enough to establish patterns of clinic visits, border crossings, and residential locations. In the post-Dobbs era, this infrastructure enables exactly the kind of cross-state surveillance abortion rights advocates warned about.

The privacy implications extend beyond individual queries. Data retention policies store vehicle sightings for 30 days as standard, longer if flagged for investigation. This means a single warrantless search for abortion-related activity creates a permanent investigative file that can be accessed repeatedly and shared across agencies. The scale is staggering: one query hitting 83,000 cameras generates millions of potential data points about reproductive healthcare access.

Surveillance of Protesters and Free Speech Chilling

Flock's ALPR network has become a tool for monitoring political dissent and protest activity. EFF's 2025 review documented law enforcement tracking protesters at events like No Kings rallies, with queries conducted without any criminal predicate. U.S. Border Patrol agents used the system to monitor streets during demonstrations, creating detailed records of First Amendment-protected activity.

This represents classic "surveillance creep"—technology initially justified for crime-fighting that expands to monitor constitutionally protected speech. The network's architecture enables out-of-jurisdiction monitoring, allowing agencies in one state to track activists attending protests hundreds of miles away. This creates a chilling effect on political participation, as individuals become aware their movements are logged and searchable.

Some jurisdictions have attempted to prevent political surveillance. Oakland's ALPR policy explicitly bans using the technology to track protesters or monitor First Amendment activities. Yet Flock's nationwide sharing model undermines these local protections—an Oakland resident attending a protest in Texas could be tracked by Texas agencies and that data shared back to California law enforcement through the network.

The implications for free speech rights are profound. Protesters at demonstrations ranging from reproductive rights rallies to immigration policy events now assume their vehicle movements are being cataloged and stored. This surveillance infrastructure doesn't require court approval, supervisory review, or even articulable suspicion of criminal activity—just an officer with network access and a search term.

Border Patrol and Unauthorized Federal Access

Immigration enforcement agencies have exploited multiple pathways into Flock's supposedly local surveillance networks. At least eight Washington state agencies enabled direct 1:1 data sharing with U.S. Border Patrol in 2025, while ten agencies experienced "back door" unauthorized access between May and August 2025. These findings, documented by EFF and the University of Washington Center for Human Rights, triggered federal investigations.

The access patterns break down into three distinct categories:

These access patterns pose severe risks to migrant protections and sanctuary policies. Local agencies in jurisdictions that prohibit immigration enforcement cooperation found their ALPR data flowing to federal agencies anyway. A Joplin officer was fired for improper use of the system, demonstrating that individual violations occur even within departments attempting oversight.

"Side door" searches represent perhaps the most insidious pathway. Officers in agencies without direct ICE relationships conducted searches using keywords like "ICE" and "immigration," effectively creating an informal information-sharing network that bypasses official policies. These queries leave audit trails showing systematic collaboration despite local prohibitions.

Flock Safety's ALPR Network Scale and Capabilities

Flock Safety has deployed automated license plate readers across more than 5,000 U.S. communities, with well over 100,000 cameras installed nationwide—roughly one camera per every 4,000 American citizens. Each camera captures license plates, timestamps, vehicle color/make/model, and distinguishing features, with no set limit on how many times a single vehicle can be tracked. The company stores this data for 30 days as standard, extending retention indefinitely for flagged investigations.

Flock claims impressive crime-solving statistics. According to their 2023 survey of 123 agencies, the technology helps solve over 700,000 crimes annually, representing approximately 10% of all reported U.S. crime. Their multilinear regression analysis from April-June 2023, correlated with FBI crime data, identifies several factors that boost clearance rates:

Real-world examples support these claims. Texas agencies solved car break-in rings by tracking suspect vehicles across jurisdictions. The "Coroner Affair" murder investigation used ALPR evidence to establish suspect timelines. Flock's methodology excludes minor offenses, focusing on substantive crimes where vehicle tracking provides investigative value.

Yet this same scale creates enormous potential for abuse. A network spanning 100,000+ cameras means a single search query can generate millions of vehicle sighting records instantly. Each camera tracks vehicles repeatedly with no upper limit, creating comprehensive movement profiles of American