Written by 
Flock Safety, founded in 2017 and headquartered in Atlanta, Georgia, operates what has become the largest automated license plate reader network in the United States, with more than 120,000 cameras deployed across 49 states, scanning approximately 20 billion license plates per month. The company is valued at $8.4 billion as of April 2026 and has raised $858 million in total venture funding. Its cameras are used by more than 5,000 law enforcement agencies and 12,000 total customers.
Despite those scale numbers, documented evidence shows that the system generates false alerts at alarming rates, and that innocent people have been pulled from their vehicles at gunpoint, mauled by police dogs, jailed for weeks, and subjected to wrongful arrest. Ohio is home to some of the most significant and legally documented incidents, and the failures extend coast to coast.
This report documents every publicly known incident of mistaken identity, false arrest, or physical harm linked to Flock cameras, beginning in Ohio and expanding nationally. The findings are drawn from court records, police body camera footage, civil lawsuit filings, internal police department communications, public records, and investigative reporting by journalists across the country.
|
Key Statistics at a Glance 120,000+ cameras deployed across 49 states 20 billion license plate reads per month $8.4 billion company valuation as of April 2026 $858 million total venture funding raised 5,000+ law enforcement agency customers 71% misread rate for stolen or felony vehicle alerts documented by Roseville, California police over 2 years 20+ publicly documented incidents of mistaken identity, false arrest, or physical harm attributed to the system nationwide |
INTRODUCTION: FLOCK SAFETY AND THE RISE OF AUTOMATED SURVEILLANCE
Flock Safety was founded in 2017 and positioned itself as a technology driven public safety company. Its core product is a stationary automated license plate reader camera branded as the Flock Falcon. The cameras use artificial intelligence to capture still images of passing vehicles, log the license plate number, and compare that number against a shared database known as the “Hot List.” The Hot List is populated by law enforcement agencies entering plates associated with stolen vehicles, wanted persons, active warrants, missing persons investigations, and felony crimes. When a plate matches, the system issues an alert to officers in real time.
The company grew rapidly through a business model that bundled hardware, software, installation, and maintenance into annual subscription contracts priced at approximately $2,400 per camera. It attracted funding from some of the most prominent venture capital firms in the country, including Andreessen Horowitz, Tiger Global, Founders Fund, Kleiner Perkins, and Y Combinator. By March 2025, Flock reported $300 million in annual recurring revenue, representing a 70 percent year over year increase. In April 2026, the company completed a funding round that set its valuation at $8.4 billion.
Flock’s stated mission is to eliminate crime and make communities safer. Its marketing materials describe the cameras as a tool to recover stolen vehicles, locate missing persons, and solve violent crimes. Law enforcement agencies from rural townships to major cities have signed contracts. State agencies in Ohio and across the country have followed. So have universities, homeowners associations, and private businesses.
What the marketing does not emphasize is what happens when the system is wrong. And the evidence shows it is wrong, often, with consequences that range from a frightening traffic stop to weeks of wrongful imprisonment and physical injury inflicted by police dogs.
OHIO: DOCUMENTED INCIDENTS OF MISTAKEN IDENTITY AND HARM
Ohio is one of the most thoroughly documented states when it comes to Flock camera failures. The incidents in Ohio span the system’s core functions, including misread license plates, stale Hot List entries, and data shared without authorization with outside agencies.
TOLEDO, OHIO, APRIL 2024: THE BRANDON UPCHURCH INCIDENT
The most extensively documented physical harm case in Ohio, and 1 of the most severe in the entire country, occurred in Toledo in April 2024. Brandon Upchurch was driving his own truck when he was pulled over by a Toledo Police Department officer. The stop was triggered by an alert from a Flock camera that misread the number “7” on Upchurch’s license plate as the number “2,” causing the system to match his plate to a stolen vehicle on the Hot List.
The responding officer, identified in reporting by Business Insider, stated on body camera footage that he believed Upchurch had tripped and was about to run as Upchurch exited the vehicle. A police dog was deployed. Upchurch was mauled. He was transported to a hospital for his injuries and then transferred to the county jail, where he was held for several hours. The charges brought against him were obstructing official business and resisting arrest. Both charges were later dismissed.
Police records reviewed by Business Insider confirmed what the body camera footage had already suggested. The officer who made the stop could be heard telling a colleague after the arrest that he believed the camera had produced a “mis-hit.” Upchurch was the registered owner of the truck. The plates were not stolen. The Flock camera had simply misread 1 character on his plate, and the resulting alert sent armed officers to his vehicle with a dog.
A civil lawsuit was filed. Upchurch settled for $35,000. The settlement did not require any admission of wrongdoing.
|
Failure Type: Single character misread (“7” read as “2”) Outcome: Police dog deployment, physical injury, wrongful arrest, charges dismissed Settlement: $35,000, no admission of wrongdoing |
JACKSON TOWNSHIP, OHIO, MARCH 2023: THE MICHAEL SMITH INCIDENT
In February 2023, the Cleveland Police Department opened a missing juvenile investigation. As part of that investigation, coordination among multiple law enforcement agencies resulted in the entry of Michael Smith’s license plate number and vehicle description into the Flock Hot List. The missing juvenile was located at an area hospital on February 23, 2023, and the Cleveland Police Department officially terminated the investigation.
2 weeks later, a Flock camera in Jackson Township, located in Stark County, captured Smith’s vehicle and matched it against the Hot List, which had not been cleared. The system issued a notification. Jackson Township police initiated a traffic stop of Smith’s vehicle. Officers contacted the Cleveland Police Department during the stop and learned that the missing juvenile had been found and that the case was closed. Smith was allowed to leave the scene. Police records show he was cleared by 5:18 a.m. The stop lasted less than 1 hour.
Smith filed suit against Jackson Township and the 4 officers who conducted the stop. That case settled on November 8, 2023. 5 days later, Smith filed a separate federal lawsuit against Flock Safety directly, case number 5:23-CV-2198 in the Northern District of Ohio. He alleged product liability, failure to warn, nonconformance with representations, and defamation, among other counts. The court ultimately dismissed his claims against Flock, finding in its April 2024 opinion that Flock did not generate the underlying Hot List information and therefore could not be held liable as a state actor or for the accuracy of data it did not create. The case nonetheless stands as 1 of the first federal lawsuits in Ohio directly naming Flock Safety as a defendant.
|
Failure Type: Stale Hot List entry, investigation closed 2 weeks prior Outcome: Wrongful stop, civil lawsuit against Jackson Township and 4 officers, separate federal suit against Flock Safety Court Finding: Flock dismissed as defendant; company did not generate the underlying Hot List data |
DAYTON, OHIO, 2026: THE DATA SHARING SCANDAL
The most consequential institutional failure involving Flock in Ohio did not involve a misread plate. It involved what Dayton city officials called “egregious violations of policy” in the sharing of Flock camera data with outside law enforcement agencies, including federal immigration enforcement.
In May 2026, the Dayton Police Department announced that it was indefinitely suspending the use of its Flock automated license plate readers pending an independent review. An internal investigation had determined that a former commander of the Support Services Division failed to implement required safeguards restricting which agencies could access data collected by the cameras, despite a city ordinance passed in 2021 that placed strict limits on how the information could be used.
The partial data report released by the Dayton commissioners office covered searches dating back to July 2023. It showed that immigration related searches had been conducted through the Dayton Flock system 7,100 times. Those searches came from 140 agencies outside Dayton’s jurisdiction. Federal Border Patrol conducted 3,703 of those searches. Lowndes County, Georgia, conducted 779. The word “ICE” was searched 1,778 times. The word “immigration” was searched 443 times, sometimes combined with words including “investigation,” “enforcement,” “violation,” or “status.” Ohio agencies also accessed the system from Wooster, Warren, and from Montgomery, Franklin, and Butler counties.
Dayton City Manager Shelley Dickstein described the violations as egregious. Dayton Mayor Shenise Turner-Sloss and Commissioner Darryl Fairchild demanded a full accounting of all audit logs from the inception of the program, beginning with a pilot program launched in 2020. They demanded an explanation of why city leadership was not informed when the breach was first identified in October 2025, months before it was disclosed publicly. The commander responsible for implementing the required safeguards no longer worked for the department by the time of the announcement.
The city had previously passed a surveillance impact ordinance requiring approval before any new technology could be acquired. Community groups had warned at the time of the initial Flock contract that proper safeguards were not in place. Those warnings proved accurate.
|
Searching Agency |
Number of Searches |
Notes |
|---|---|---|
|
Federal Border Patrol |
3,703 |
Prohibited under Dayton city ordinance |
|
Lowndes County, Georgia |
779 |
Out of state, no authorization |
|
Searches containing “ICE” |
1,778 |
Immigration enforcement related |
|
Searches containing “immigration” |
443 |
Combined with enforcement, violation, or status |
|
Total immigration related searches |
7,100 |
From 140 agencies outside Dayton jurisdiction |
CUYAHOGA COUNTY AND STATE CONTRACTS: THE OHIO SPENDING RECORD
Ohio’s commitment to Flock technology extends far beyond Toledo and Dayton. Cleveland City Council voted 9 to 6 on July 15, 2026, to approve a 6 month extension of its Flock contract valued at $125,000. Cleveland had installed 100 Flock Falcon cameras across the city in 2023 under an initial $250,000 contract. Mayor Justin Bibb’s administration had sought a 1 year, $250,000 renewal. The council shortened the term amid debates over privacy and data sharing, and attached conditions including a bar on sharing Cleveland Flock data with the Northeast Ohio Regional Fusion Center without a valid search warrant, a requirement for a public transparency portal, quarterly reports to council, and a contract penalty of up to 25 percent for privacy or data sharing violations.
Shaker Heights, a city neighboring Cleveland, approved a Flock contract with a maximum value of $204,550 over 5 years. A police memorandum projected combined costs of $54,650 in 2025 and $50,500 per year in 2026, 2027, and 2028. Strongsville authorized an initial 24 month agreement for 19 Flock readers in January 2024, at $72,250 for the first year and $63,500 annually, for an initial contract total of $135,750. The city later authorized a 6 reader expansion adding an additional $37,150 in the first year and $33,000 annually, bringing the maximum contract value to $204,550.
Reporting by Signal Cleveland identified more than 1,700 license plate readers across Cuyahoga County alone, including government and privately owned systems.
At the state level, Ohio agencies responsible for workers’ compensation, fire investigations, adult prisons, juvenile prisons, and organized crime investigations have all contracted with Flock Safety. Disclosed spending totals approximately $620,500 across those agencies, covering different time periods and purposes. The Bureau of Workers’ Compensation paid Flock $24,000 in the most recently reported year and the same amount in 2025. Adult corrections contracts totaled approximately $384,500. The attorney general’s Organized Crime Investigations Commission paid approximately $105,000. Officials at multiple state agencies declined to provide specifics on why they access the Flock network or how data is used.
Ohio State University signed its 3rd agreement with Flock Safety in July 2026, a 2 year contract totaling $378,000, representing a 20 percent annual increase from the previous deal. Ohio State’s 63 cameras on and around its Columbus campus have been in place since a free pilot program in 2022. Ohio State officials said the cameras have been used to “solve crimes and assist law enforcement” but acknowledged they have no formal data tracking to support that claim. Ohio State’s 3 agreements with Flock Safety total more than $1 million through 2028.
Case Western Reserve University and other Ohio colleges have also partnered with Flock Safety. Colerain Township in Hamilton County installed Flock license plate readers in mid 2025 and became the subject of a public records dispute before the Ohio Supreme Court in 2026, as citizens sought access to the terms of the Flock data arrangement. The court ultimately denied the records request on procedural grounds without resolving the merits of public access to the data.
|
Ohio Entity |
Contract / Spending |
Notes |
|---|---|---|
|
Cleveland (6 month extension) |
$125,000 |
Council approved 9 to 6, July 15, 2026; 100 cameras |
|
Shaker Heights (5 year max) |
$204,550 |
$54,650 in 2025; $50,500 per year 2026 to 2028 |
|
Strongsville (initial + expansion) |
$204,550 |
19 readers initially; 6 reader expansion added |
|
Adult Corrections contracts |
~$384,500 |
Multiple time periods |
|
AG Organized Crime Investigations Commission |
~$105,000 |
Specific use details not disclosed |
|
Bureau of Workers’ Compensation |
$24,000 (most recent year) |
Same amount paid in 2025 |
|
Ohio State University (3rd contract) |
$378,000 |
2 year term; 63 cameras; total spending exceeds $1 million |
|
State agency total (disclosed) |
~$620,500 |
Across workers’ compensation, corrections, AG, juvenile prisons, fire investigations |
NATIONAL INCIDENTS: A DOCUMENTED RECORD OF FAILURES
The following section presents every publicly documented incident of mistaken identity, false arrest, or physical harm linked to Flock Safety cameras outside of Ohio, organized chronologically.
ATHERTON, CALIFORNIA, APRIL 2021
Jason Burkleo was driving through the Silicon Valley community of Atherton when he was pulled over at gunpoint by local police, instructed to lie on the ground, and handcuffed. A Flock camera had confused the letter “H” on his license plate with the letter “M.” Officers noted in their reports that “the Flock cameras misread” the plate. Dirt on the plate had obscured 1 character, causing the misread. Burkleo filed a civil lawsuit. He settled with the town and the police officers for $45,000. The Atherton Police Department denied any wrongdoing and maintained that it acted properly and lawfully.
|
Failure Type: Character misread (“H” read as “M”); plate obscured by dirt Outcome: Detained at gunpoint, handcuffed, civil lawsuit filed Settlement: $45,000 |
SEDGWICK COUNTY, KANSAS, OCTOBER 2023
A Flock camera failed to distinguish a temporary paper plate from a permanent license plate. This reading error caused deputies to detain an innocent couple at gunpoint. No charges were filed.
DETROIT, MICHIGAN, SEPTEMBER 2023
Officers acted on misread plate data and detained an innocent woman at gunpoint. Her autistic child was placed in a squad car during the stop. Her vehicle was impounded and held for weeks before being returned. No charges resulted.
HOUSTON, TEXAS, AUGUST 2023
Officers misread a stolen vehicle alert and detained 2 innocent men at gunpoint. The error stemmed from plate reader data produced by the Flock system. No charges were filed.
ESPANOLA, NEW MEXICO, JULY 2023
In 1 incident that month, a Flock camera misread a digit and led to 2 sisters being detained at gunpoint. In a separate incident days later in the same city, officers entered incorrect plate data themselves into the system and detained an innocent teenager at gunpoint. Both incidents occurred within weeks of each other and involved different failure mechanisms, 1 originating with the camera software and 1 originating with a human data entry error that Flock’s system then dutifully matched and alerted on.
FAYETTEVILLE, NORTH CAROLINA, JULY 2023
Officers acted on Flock plate reader data and detained an innocent woman at gunpoint. The incident stemmed from a misread by the system. No charges were filed.
YORK COUNTY, SOUTH CAROLINA, FEBRUARY 2024
An innocent motorist was stopped at gunpoint after a Flock camera misread a single character on his license plate. The York County Sheriff’s Office attributed part of the problem to a tinted plate cover. No charges were filed.
NORTH MYRTLE BEACH, SOUTH CAROLINA, MAY 2024
Officers acting on a Flock alert misidentified a vehicle and detained an innocent teenager at gunpoint. The teenager was not the subject of any investigation. No charges were filed.
LAFAYETTE, INDIANA, JULY 2024
A Flock camera misread a plate, triggering an alert that sent officers to detain an innocent driver at gunpoint. No charges were filed.
MORRISTOWN, TENNESSEE, JUNE 2024
A Flock camera read the letter “O” as the number “0.” The alert sent officers to stop a vehicle carrying 2 grandparents and their 3 year old granddaughter. All 3 were held at gunpoint. The grandparents were not suspects in any crime. No charges were filed.
SAN DIEGO, CALIFORNIA, NOVEMBER 2025
Officers investigating an attempted carjacking relied on Flock’s “Vehicle Fingerprint” artificial intelligence feature, which allows matching based on vehicle appearance characteristics including color, make, decals, and body type, even without a plate match. Officers flagged a red Alfa Romeo captured by a Flock camera 5 miles from the crime scene as the suspect vehicle. They arrested the occupants, Hugo Parra and Ariel Beltran.
Timestamp data embedded within the Flock system itself showed that the captured vehicle could not have been at the crime scene at the time of the incident. That metadata was not consulted before the arrest. A witness misidentification compounded the error. Parra was wearing a white hoodie at the time of arrest. The suspect in the carjacking had been described as wearing gray. Parra, who was on probation at the time, was jailed for nearly 30 days before being released.
A civil rights lawsuit was filed in June 2026. Parra and Beltran are each seeking $1.5 million in damages. The lawsuit alleges negligence and civil rights violations and argues that San Diego Police dismissed cellphone location data that corroborated Parra’s alibi. San Diego spends $2 million annually to maintain access to the Flock network.
|
Failure Type: Vehicle Fingerprint AI feature misidentification; Flock timestamp data showing impossibility of match not reviewed before arrest Outcome: 2 men arrested; 1 jailed nearly 30 days; civil rights lawsuit filed June 2026 Damages Sought: $1.5 million each; San Diego annual Flock cost: $2 million |
BOULDER COUNTY, COLORADO, DECEMBER 2025
A Colorado driver was repeatedly stopped after her own license plate was mistakenly added to a Flock hotlist by officers. The error was not corrected for an extended period. Each subsequent stop was initiated by a valid Flock alert generated by incorrect human data entry.
REDMOND, WASHINGTON, AUGUST 2025
A data linking error inside the Flock system connected a suspect’s vehicle to his innocent father’s record. Officers detained the father, the wrong man, based on the alert.
SHERWOOD, ARKANSAS, FEBRUARY 2026
A Flock camera misread a single digit on an innocent couple’s license plate and flagged the plate as stolen. Officers responded and ordered the couple out of their vehicle at gunpoint. A 6 week old baby was left alone in the back seat during the stop. An officer who responded to questions about the incident described the cameras as not “completely perfect.” No charges were filed against the couple.
CHERRY HILLS, COLORADO, APRIL 2026
A driver was stopped repeatedly after officers mistakenly entered his own license plate number into a Flock hotlist instead of a suspect’s plate. The driver had no connection to any criminal matter. Each subsequent stop was generated by the Flock system correctly matching a plate that had been incorrectly placed on the hotlist by a human officer.
VOLUSIA COUNTY, FLORIDA, APRIL 2026
A Flock camera capture led officers to misidentify a driver’s vehicle as 1 linked to a fatal accident. The driver was arrested and spent 13 days in jail before the error was discovered and charges were dropped.
YORK COUNTY, SOUTH CAROLINA, JULY 2026
A second incident in York County occurred approximately 2 years after the first. A driver identified as Steven Melvin was stopped at gunpoint after a Flock camera misread a single letter on his plate. Melvin filed a lawsuit. The case remains active.
PLYMOUTH, MINNESOTA, JUNE 2026
Automotive journalist Joel Feder was tracked by law enforcement for 2 consecutive days before being boxed in by 4 squad cars outside a retail store while driving a loaner Range Rover Sport. The alert traced back to a stolen plate report entered into the National Crime Information Center database in California. That report was missing 2 middle digits, which were printed smaller than the rest of the plate. Flock’s system matched on the shortened character sequence. The responding officers never checked the full plate against the actual photograph produced by Flock’s own camera before moving in.
SCOTTS BLUFF, NEBRASKA, JUNE 2026
Within 1 week of the Plymouth, Minnesota incident, a second automotive journalist driving a similarly configured loaner Range Rover was pulled over in Nebraska based on the same underlying stolen plate report. He was detained for approximately 1 hour before being released.
THE ACCURACY PROBLEM: WHAT THE NUMBERS REVEAL
The incidents listed in this report are not isolated technical glitches. They reflect a systemic and documented pattern in Flock’s error rate. The numbers available from independent audits and internal police department records tell a story the company’s marketing does not.
In a 2 year analysis covering 2023 and 2024, the Roseville Police Department in California examined 1,427 alerts received from Flock cameras during that period. In 71 percent of the cases where a vehicle was flagged as stolen or connected to a felony, the system had misread the license plate. The error rate for that alert category was 71 percent over 1,427 alerts. Roseville’s own crime center supervisor, Amanda Jones, wrote to a Flock senior product manager in January 2026 stating that trust in the system was degrading “each time we run into this major issue.” A Roseville lieutenant reviewed a report showing 16 incorrect plates in a single week and responded with a single word: “Ouch.”
Flock disputed the characterization. A company spokesperson told journalists that most of the misread data was from 2025 or before, and that camera performance had significantly improved. Roseville police disputed that claim. Flock also noted that Roseville used an older hardware configuration and non-standard installation locations, with cameras placed further from vehicles and higher than recommended, and that the city had requested that cameras scan only the backs of vehicles to avoid capturing faces.
The independent research firm IPVM tested Flock cameras in 2021 and concluded that the system misidentified which state a license plate was from in approximately 1 in 10 reads, and that Flock regularly misclassified vehicle type and make. IPVM stated that Flock subsequently blocked it from purchasing cameras for further testing.
The Electronic Frontier Foundation published an analysis in November 2024 finding that automated license plate reader systems misread the state on a license plate in approximately 1 in 10 cases. Independent police research published in The Police Chief magazine found that approximately 37 percent of fixed position automated license plate reader hits are erroneous. Flock’s own marketing claims accuracy rates in “the high 90s,” meaning the company itself acknowledges a base error rate of at least 7 percent per scan. At 20 billion plate reads per month, a 1 percent error rate produces approximately 200 million misreads every single month.
|
Documented Character Confusion Errors The most common character confusions documented in police records and company correspondence include the following pairs: “N” misread as “V” “1” misread as “4” “3” misread as “2” “7” misread as “2” “O” misread as “0” “H” misread as “M” Additional failures stem from plates obscured by dirt, license plate frames, tree branches, or distance from the camera. Hot List entries that remain active after investigations close produce additional false alerts that are entirely human in origin but amplified by the Flock network across jurisdictions. |
Max Isaacs, director of technology law and policy with the Policing Project at New York University’s School of Law, described Roseville’s misread data as “stunning” and stated that a 71 percent misread rate is “really beyond the pale” and that “in the absence of concrete rules that directly address the problem of misreads, this technology is not safe to be deployed in communities.”
THE CIVIL LIBERTIES DIMENSIONS
The physical harm and wrongful arrest cases represent only part of the documented concern surrounding Flock Safety. Civil liberties organizations, legal scholars, and elected officials have raised alarms about the broader surveillance architecture the company has built.
The ACLU has characterized Flock’s network not as a localized investigative tool but as a searchable, cross jurisdictional movement history database. Because camera data is shared across Flock’s network, a single query can reach data from thousands of agencies simultaneously. The EFF documented in 2025 that 1 query to the Flock system accessed 83,345 cameras across nearly the entire country. That query was conducted by a sheriff’s department in Johnson County, Kansas, in the context of what police characterized as a missing person investigation but was actually an abortion case. A Johnson County official ran 2 searches, 1 with the note “had an abortion, search for female,” probing 6,809 separate networks in the 2nd search.
The EFF’s 2025 annual review documented the system being used to track protesters, monitor specific communities, and surveil women seeking reproductive healthcare. In Georgia, multiple officers faced criminal charges for using automated license plate reader databases to track former romantic partners and personal acquaintances. U.S. Representative Raja Krishnamoorthi and U.S. Representative Robert Garcia launched a formal congressional investigation into Flock’s role in enabling what they described as invasive surveillance that threatens the privacy, safety, and civil liberties of women, immigrants, and other vulnerable Americans.
Illinois Secretary of State Alexi Giannoulias launched an audit after EFF research showed Flock allowed U.S. Customs and Border Protection to access Illinois data in violation of state privacy laws. In August 2025, Flock Safety acknowledged that it had engaged in limited pilot programs with U.S. Customs and Border Protection and Homeland Security Investigations and announced it was halting those programs.
In Dayton, Ohio, as detailed earlier in this report, federal immigration enforcement agencies conducted 7,100 immigration related searches through the city’s Flock system, including 3,703 by Border Patrol, in direct violation of the city’s own ordinance prohibiting that use.
The ACLU’s Chad Marlow described Flock as “a company that trades privacy for profit,” arguing that because the system records every vehicle that passes any Flock camera, “we have a company working to record the comings and goings of every single person in the country,” an outcome he called a “very un-American take on privacy.”
The West Virginia Freedom Caucus sought to ban Flock cameras statewide, with chairman Chris Anders stating that the cameras represent “a dramatic expansion of government surveillance that is fundamentally at odds with the Fourth Amendment.” That legislation, called the Fourth Amendment Restoration Act, was introduced for 2 consecutive years and failed to reach the governor’s desk in both attempts.
Community oversight boards in at least 1 city concluded that cameras played no meaningful role in local crime investigations since installation in 2022. That city ended its Flock contract in August 2025. Santa Cruz, California, and Los Altos Hills, California, also ended their Flock contracts amid public pressure.
LEGAL CHALLENGES AND COURT PROCEEDINGS
The legal landscape surrounding Flock Safety is evolving rapidly, and courts have reached differing conclusions depending on the specific facts before them.
In Norfolk, Virginia, plaintiffs Lee Schmidt and Crystal Arrington sued the city and its police department, arguing that Norfolk’s network of 176 Flock cameras, organized into 75 camera clusters, constituted a Fourth Amendment search and violated their privacy rights. Federal District Judge Mark S. Davis granted summary judgment to the City of Norfolk in January 2026, finding that the system did not capture enough images to reconstruct the whole of any individual’s movements. The judge applied precedent holding that fixed position cameras photographing license plates on public roads do not infringe on a reasonable expectation of privacy. Flock Safety’s chief legal officer stated that the ruling “aligns with strong national precedent, as over 30 state and federal courts across the country have concluded that fixed location automated license plate readers do not infringe on an individual’s reasonable expectation of privacy.”
The plaintiffs announced they would appeal to the Fourth Circuit Court of Appeals. The Institute for Justice, which represented the plaintiffs, stated that “as abuses of these automated license plate reader systems are mounting nationwide, it’s more important than ever to vindicate the people’s right to security from mass surveillance guaranteed by the Fourth Amendment.”
In California, plaintiffs Daniel Javorsky and Anthony Mayor filed a class action lawsuit against Flock Group, Inc. in 2026, alleging violations of California’s Automated License Plate Reader Privacy Act, California’s Unfair Competition Law, and the California Constitution. The complaint accuses Flock of building an “Orwellian mass surveillance infrastructure that is practically impossible to avoid” and of sharing data on California drivers with out of state and federal law enforcement agencies in violation of state law. The lawsuit notes that Flock has repeatedly and publicly disclaimed responsibility for those violations, instead pointing to its customers. The plaintiffs argue that Flock cannot rely on what the complaint calls “weaponized incompetence” when its legal obligations are clear.
California Attorney General Rob Bonta sued the city of El Cajon in October 2023 for illegally sharing Flock license plate data with out of state law enforcement in violation of the Automated License Plate Reader Privacy Act.
The EFF and the ACLU of Northern California filed a lawsuit in November 2025 against the city of San Jose and its police department over their use of the Flock system.
In Ohio’s Northern District federal court, the Smith v. Flock Safety decision of April 2024 dismissed the direct claims against Flock while confirming the factual record of the Jackson Township stop. The court found that Flock did not generate the Hot List entry and therefore could not be held liable under product liability or constitutional theories for the accuracy of data it matched but did not create.
|
Case / Jurisdiction |
Year |
Outcome / Status |
|---|---|---|
|
Smith v. Flock Safety, N.D. Ohio (5:23-CV-2198) |
April 2024 |
Flock dismissed; court found Flock did not create Hot List data |
|
Schmidt and Arrington v. City of Norfolk, Virginia |
January 2026 |
Summary judgment for city; Fourth Amendment challenge rejected; appeal announced to Fourth Circuit |
|
Javorsky and Mayor v. Flock Group, Inc. (California class action) |
2026 |
Active; alleges violations of California Automated License Plate Reader Privacy Act and state constitution |
|
California AG v. City of El Cajon |
October 2023 |
Sued for illegal out of state data sharing |
|
EFF and ACLU v. City of San Jose |
November 2025 |
Active civil suit over Flock system use |
|
Parra and Beltran v. San Diego (civil rights) |
June 2026 |
Active; $1.5 million sought per plaintiff |
|
Melvin v. York County (South Carolina) |
2026 |
Active lawsuit; single letter misread |
CORPORATE RESPONSE AND INDUSTRY ACCOUNTABILITY
Flock Safety has consistently defended its technology while acknowledging that errors occur. In response to documented incidents, company representatives have pointed to installation conditions, older hardware, human error by law enforcement agencies, and the responsibility of individual officers to independently verify alerts before taking enforcement action.
The company has announced feature updates in response to public criticism. The EFF stated in 2025 that those updates miss the point because the core problem is not any specific feature but the architecture of a nationwide interconnected surveillance network that creates risks no software update can eliminate.
In June 2026, Flock Safety reached an $8.4 billion valuation following a $200 million funding round, even as protests spread across the country and communities debated canceling contracts. The company crossed $300 million in annual recurring revenue in early 2025. Its investor base includes Andreessen Horowitz, Founders Fund, Kleiner Perkins, Meritech Capital, Tiger Global, Greenoaks Capital, and Y Combinator.
Amazon’s Ring announced it would end its partnership with Flock following public backlash over a Super Bowl advertisement for Ring cameras. Flock described the decision as mutual.
U.S. Senator Alex Padilla of California stated that inaccuracies stemming from law enforcement’s use of Flock technology are “deeply alarming.”
CONCLUSIONS AND FINDINGS
The documented record of Flock Safety camera failures is not a collection of isolated incidents. It is a pattern. In every documented case in this report, at least 1 person was stopped at gunpoint, placed in handcuffs, bitten by a police dog, jailed for hours or weeks, or otherwise subjected to law enforcement force as a direct result of a Flock camera alert that was either wrong or based on stale or incorrectly entered data.
In total, more than 20 documented incidents of mistaken identity, false arrest, or physical harm have been attributed to the Flock system across the country. Independent experts and law enforcement officials have stated that the publicly known cases almost certainly represent only a fraction of the total. No federal agency currently tracks or publicly reports wrongful stops caused by automated license plate reader errors. No mandatory reporting mechanism exists.
In Ohio alone, 1 man was physically attacked by a police dog and jailed. 1 man was stopped on an investigation that had been closed for 2 weeks. The state’s 2nd largest city, Dayton, allowed thousands of immigration related searches to be conducted through its Flock system in direct violation of its own ordinance. The state’s flagship university has spent more than $1 million on a technology whose crime solving record university officials acknowledged they cannot document.
The Roseville, California data is the most precise window available into the scale of the problem. A 71 percent misread rate over 1,427 alerts in 2 years is not an outlier. It is a documented operating condition of a system that generates 20 billion reads per month and is deployed across 49 states with 120,000 cameras. The math is not ambiguous. At even a 1 percent error rate, 200 million misreads occur each month. Most will not result in a stop. But the ones that do carry guns, dogs, handcuffs, and the power to detain.
The technology will continue to expand as long as contracts continue to be signed and as long as oversight mechanisms remain weak or nonexistent. The people who pay the price when the system is wrong are not the company’s shareholders. They are the drivers pulled from their vehicles at gunpoint because a camera confused a “7” with a “2.”
|
Summary of Core Findings 1. Flock Safety operates the largest automated license plate reader network in the United States, with 120,000+ cameras scanning 20 billion plates per month. 2. The system’s documented misread rate in at least 1 independently audited police department reached 71 percent for stolen or felony vehicle alerts over a 2 year period. 3. At least 20 publicly documented cases of wrongful stops, false arrests, or physical harm have been directly attributed to Flock camera errors across the country. 4. In Ohio, documented incidents include 1 man attacked by a police dog, 1 man stopped on a closed investigation, and thousands of unauthorized immigration searches conducted through Dayton’s system in violation of city ordinance. 5. No federal mandatory reporting system exists for wrongful stops caused by automated license plate reader errors. The publicly documented cases almost certainly undercount the true total. 6. As of August 1, 2026, Flock Safety is valued at $8.4 billion and continues to expand its network, while legal challenges in multiple jurisdictions remain unresolved. |
This report is current as of August 1, 2026. All facts, figures, case citations, and incident descriptions are drawn from court records, police reports, civil litigation filings, body camera footage records, public records disclosures, and published investigative reporting. This report does not constitute legal advice.