Flock license plate readers: Difference between revisions

(5 intermediate revisions by 5 users not shown)

Line 24:

Data collected includes location history that can reveal sensitive information about medical visits, religious attendance, political activities, and personal associations. While Flock states data is deleted after 30 days, contracts grant them ''"perpetual, worldwide, royalty-free license"'' to use anonymized data indefinitely.<ref>{{Cite web|url=https://www.flocksafety.com/terms-and-conditions|title=Terms and Conditions|work=Flock Safety|access-date=2025-08-23}}</ref> The system shares data across a network of over 4,800 law enforcement agencies nationally.<ref>{{Cite web|url=https://www.404media.co/lawsuit-argues-warrantless-use-of-flock-surveillance-cameras-is-unconstitutional/|title=Lawsuit Argues Warrantless Use of Flock Surveillance Cameras Is Unconstitutional|work=404 Media|date=2024|access-date=2025-08-23}}</ref>

====="Anonymized ~~data~~"=====

====="Anonymized Data"=====

While Flock ~~defines anonymized data~~ as customer data that is ''"permanently stripped of identifying details and any potential personally identifiable information"'' and is rendered so that a person or entity ''"can no longer be identified directly or indirectly,"'' this definition includes information such as vehicle make, model, color, location patterns, and other non–license-plate attributes.<ref>{{cite web |date=2025-07-22 |title=Terms and Conditions |url=https://www.flocksafety.com/legal/terms-and-conditions |access-date=2025-08-23 |website=Flock Safety}}</ref>

While Flock's Terms and Conditions define "Anonymized Data" as customer data that is ''"permanently stripped of identifying details and any potential personally identifiable information"'' and is rendered so that a person or entity ''"can no longer be identified directly or indirectly,"'' this definition includes information such as vehicle make, model, color, location patterns, and other non–license-plate attributes.<ref>{{cite web |date=2025-07-22 |title=Terms and Conditions |url=https://www.flocksafety.com/legal/terms-and-conditions |access-date=2025-08-23 |website=Flock Safety}}</ref>

Privacy researchers caution that mobility datasets labeled as "anonymized" can still be re-identified. A 2013 MIT study found that just four spatio-temporal points uniquely identified 95% of individuals in an anonymized location dataset.<ref>{{cite journal |last=de Montjoye |first=Y.-A. |last2=Hidalgo |first2=C. A. |last3=Verleysen |first3=M. |last4=Blondel |first4=V. D. |year=2013 |title=Unique in the Crowd: The privacy bounds of human mobility |url=https://www.nature.com/articles/srep01376 |journal=Scientific Reports |volume=3 |pages=1376 |doi=10.1038/srep01376 |access-date=2025-08-23}}</ref> Multiple peer-reviewed studies from 2018-2024 demonstrate that "anonymized" vehicle location data can be re-identified with high accuracy. A 2022 study showed researchers could re-identify drivers from raw vehicle network data with 97% accuracy by exploiting interdependencies in sensor measurements.<ref>{{cite web|url=https://www.sciencedirect.com/science/article/abs/pii/S0167404822002139|title=Privacy-preserving vehicle trajectory matching|website=ScienceDirect|date=2022|access-date=2025-10-05}}</ref> Research published in the Journal of Computer Science and Technology (2022) found that even 3-4 location points can uniquely identify individuals.<ref>{{cite journal|url=https://link.springer.com/article/10.1007/s11390-022-2409-x|title=Experiments and Analyses of Anonymization Mechanisms for Trajectory Data Publishing|journal=Journal of Computer Science and Technology|date=2022|access-date=2025-10-05}}</ref>

Line 102:

In Illinois, a Flock representative allegedly threatened a Department of Transportation official with police pressure when questioned about permit applications.<ref>{{Cite web |last=Uprise RI Staff |date=2024-10-23 |title=As Flock Surveillance Cameras Proliferate in Rhode Island, Lawsuit Challenges Their Legality |url=https://upriseri.com/as-flock-surveillance-cameras-proliferate-in-rhode-island-lawsuit-challenges-their-legality/ |url-status=live |access-date=2025-08-25 |website=UPRISE RI}}</ref>

====='''In Evanston:'''=====

Flock was ordered to remove 18 stationary cameras. The city put the contract with Flock on a 30-day termination notice on August 26. Flock Initially appeared to comply, removing 15 of the cameras by September 8. Later, Flock was caught reinstalling all of them by the following Tuesday without authorization from the city.

The city of Evanston responded with a cease-and-desist order for Flock to remove the new and unauthorized camera equipment.

Because Flock reinstalled the cameras without permission, Evanston was forced to cover the cameras with tape and bags to block them from potentially logging vehicle data.

This situation is particularly absurd, as the city was left with little choice but to physically shield the cameras due to Flock’s unauthorized action, and outright noncompliance with the wishes of the city.

Flock appears apprehensive to comply and remove the hardware, with a representative of flock stating:

“We continue to be optimistic that we will have the opportunity to have a constructive dialogue to address the City’s concerns and resume our successful partnership making Evanston safer”

While it is clear that flock wishes to continue their contract, this does not address the wishes of the city. In this regard, A Flock spokesperson stated: “the cameras are not active and will be uninstalled”

This raises concerns, as it is unclear why flock chose to incur costs to install unauthorized inactive cameras after the initial request for removal. Additionally, it is not possible, or remains extremely difficult and costly for the city to independently verify that the ‘inactive’ cameras are truly disabled.

It remains unclear if there is concrete financial incentive for flock to partake in these possibly illegal or otherwise unethical actions.

More physical cameras are ostensibly capable of creating more data, which is collected in aggregate by flock, and selectively shared with other clients. Additional data could create a clear value proposition for certain flock services, and thus effects the desirability of the services provided by flock, and potentially, the price which flock can expect to charge for their services.<ref>{{Cite web |last=Harrison |first=Alex |date=September 25th, 2025 |title=City covers Flock cameras while waiting for removal |url=https://evanstonroundtable.com/2025/09/25/city-covers-up-flock-cameras-while-waiting-for-removal/ |url-status=live |archive-url=https://archive.ph/PD1qe |archive-date=2025-10-10 |access-date=2025-10-10 |website=Evanston Roundtable}}</ref>

===City rejections and terminations===

Line 137:

Line 158:

===State audit findings===

California State Auditor's February 2020 investigation found Los Angeles Police Department, with a 320 million image database, had no ALPR-specific policy at all. The audit found 96% of agencies claim to have policies, but most are incomplete. Data retention periods varied wildly with no justification. LAPD maintained a minimum five-year retention period, yet couldn't demonstrate that images stored for years had investigative value. The audit found that 99.9% of the 320 million images Los Angeles stores are for vehicles that were not on a hot list when the image was made.<ref>{{cite web|url=https://information.auditor.ca.gov/reports/2019-118/summary.html|title=Automated License Plate Readers|website=California State Auditor|date=2020-02|access-date=2025-10-05}}</ref>

California State Auditor's February 2020 investigation found Los Angeles Police Department, with a 320 million image database, had no ALPR-specific policy at all. The audit found 96% of agencies claim to have policies, but most are incomplete. Data retention periods varied wildly with no justification. LAPD maintained a minimum five-year retention period, yet couldn't demonstrate that images stored for years had investigative value. The audit found that 99.9% of the 320 million images Los Angeles stores are for vehicles that were not on a hot list when the image was made.<ref>{{cite web|url=https://information.auditor.ca.gov/reports/2019-118/summary.html|title=Automated License Plate Readers|website=California State Auditor|date=2020-02-13|access-date=2025-10-05}}</ref>

New Jersey provides a contrasting model with mandatory annual audits of all 523 law enforcement agencies. The 2024 audit reported only two significant violations, both involving users who hadn't completed required training.<ref>{{cite web|url=https://nj.gov/njsp/ALPR/pdf/2024_Audit_Automated_License_Plate_Recognition_(ALPR)_Data_Collected_Utilized_NJ_Law_Enforcement_Agencies.pdf|title=2024 Audit of Automated License Plate Recognition (ALPR) Data|website=New Jersey State Police|date=2024|access-date=2025-10-05}}</ref>

Line 144:

Line 165:

===Cost-benefit analysis===

Arizona Department of Transportation's 2008 study of generic ALPR technology (predating Flock Safety by nine years) estimated $9.98 million for a hypothetical statewide ALPR system. The projected benefit-to-cost ratio of 9.6:1 came entirely from registration and insurance compliance, not crime reduction.<ref>{{cite web|url=https://apps.azdot.gov/files/ADOTLibrary/publications/project_reports/PDF/AZ637.pdf|title=Automated License Plate Recognition Technology Implementation Report|website=Arizona Department of Transportation|date=2008-06|access-date=2025-10-05}}</ref>

Arizona Department of Transportation's 2008 study of generic ALPR technology (predating Flock Safety by nine years) estimated $9.98 million for a hypothetical statewide ALPR system. The projected benefit-to-cost ratio of 9.6:1 came entirely from registration and insurance compliance, not crime reduction.<ref>{{cite web|url=https://apps.azdot.gov/files/ADOTLibrary/publications/project_reports/PDF/AZ637.pdf|title=Automated License Plate Recognition Technology Implementation Report|website=Arizona Department of Transportation|date=2008-06-01|access-date=2025-10-05}}</ref>

Colorado's Office of Research and Statistics reported that while ALPR systems are expanding, independent academic research contradicts vendor claims. A 2011 George Mason University study concluded ALPRs "do not achieve a prevention or deterrent effect" on crime.<ref>{{cite web|url=https://cdpsdocs.state.co.us/ors/Docs/Briefs/2024-05_InDetail-ALPR.pdf|title=Automated License Plate Readers (In Detail)|website=Colorado Division of Criminal Justice|date=2024-05|access-date=2025-10-05}}</ref>

Colorado's Office of Research and Statistics reported that while ALPR systems are expanding, independent academic research contradicts vendor claims. A 2011 George Mason University study concluded ALPRs "do not achieve a prevention or deterrent effect" on crime.<ref>{{cite web|url=https://cdpsdocs.state.co.us/ors/Docs/Briefs/2024-05_InDetail-ALPR.pdf|title=Automated License Plate Readers (In Detail)|website=Colorado Division of Criminal Justice|date=2024-05-01|access-date=2025-10-05}}</ref>

Oakland Police Department reported 182 arrests from ALPR in the first year, representing 1.4% of homicides, robberies, burglaries, and firearm assaults. The Northern California Regional Intelligence Center states approximately 1-2 vehicles out of 1,000 initiate alerts—a hit rate of just 0.1-0.2%.<ref>{{cite web|url=https://ncric.ca.gov/wp-content/uploads/2021/10/California-Law-Enforcement-ALPR-FAQ.pdf|title=California Law Enforcement ALPR FAQ|website=NCRIC|date=2021|access-date=2025-10-05}}</ref>

Line 153:

Line 174:

The locations of many Flock Cameras have been mapped by the OpenStreetMap project. A viewer of the locations of these cameras is located here: https://deflock.me/map

==~~References~~==

==== Cease and desist to Deflock.me ====

~~{{reflist}}~~

DeFlock.me is a website allowing users to log and view the locations of ALPRs, such as Flock products. On January 30th, 2025, Flock sent a cease and desist notice to the owner of DeFlock demanding the name of the website be changed to exclude the company's brand name. The letter also stated that "the Website also implies that various license plate readers are vulnerable to security hacks. . ." which Flock alleged ". . . provides a false impression about the security of Flock Products." <ref>''[https://www.eff.org/files/2025/02/26/2025_01_31_deflock_cd_ex-3.pdf “DEFLOCK”- Cease and Desist]. Institute for Justice. 2025-01-30. Retrieved 2025-10-17.''</ref>

==External links==

Line 163:

Line 184:

*[https://wiki.alprwatch.org/index.php/Main_Page ALPR Watch Wiki]

~~[[Category:Surveillance]]~~

==References==

~~[[Category:Privacy violations]]~~

~~[[Category:Law enforcement technology]]~~

[[Category:Flock License Plate Readers]]

~~[[Category:Fourth Amendment issues]]~~

[[Category:~~Data security~~]]

@@ Line 24: / Line 24: @@
 Data collected includes location history that can reveal sensitive information about medical visits, religious attendance, political activities, and personal associations. While Flock states data is deleted after 30 days, contracts grant them ''"perpetual, worldwide, royalty-free license"'' to use anonymized data indefinitely.<ref>{{Cite web|url=https://www.flocksafety.com/terms-and-conditions|title=Terms and Conditions|work=Flock Safety|access-date=2025-08-23}}</ref> The system shares data across a network of over 4,800 law enforcement agencies nationally.<ref>{{Cite web|url=https://www.404media.co/lawsuit-argues-warrantless-use-of-flock-surveillance-cameras-is-unconstitutional/|title=Lawsuit Argues Warrantless Use of Flock Surveillance Cameras Is Unconstitutional|work=404 Media|date=2024|access-date=2025-08-23}}</ref>
-====="Anonymized data"=====
+====="Anonymized Data"=====
-While Flock defines anonymized data as customer data that is ''"permanently stripped of identifying details and any potential personally identifiable information"''  and is rendered so that a person or entity ''"can no longer be identified directly or indirectly,"'' this definition includes information such as vehicle make, model, color, location patterns, and other non–license-plate attributes.<ref>{{cite web |date=2025-07-22 |title=Terms and Conditions |url=https://www.flocksafety.com/legal/terms-and-conditions |access-date=2025-08-23 |website=Flock Safety}}</ref>
+While Flock's Terms and Conditions define "Anonymized Data" as customer data that is ''"permanently stripped of identifying details and any potential personally identifiable information"''  and is rendered so that a person or entity ''"can no longer be identified directly or indirectly,"'' this definition includes information such as vehicle make, model, color, location patterns, and other non–license-plate attributes.<ref>{{cite web |date=2025-07-22 |title=Terms and Conditions |url=https://www.flocksafety.com/legal/terms-and-conditions |access-date=2025-08-23 |website=Flock Safety}}</ref>
 Privacy researchers caution that mobility datasets labeled as "anonymized" can still be re-identified. A 2013 MIT study found that just four spatio-temporal points uniquely identified 95% of individuals in an anonymized location dataset.<ref>{{cite journal |last=de Montjoye |first=Y.-A. |last2=Hidalgo |first2=C. A. |last3=Verleysen |first3=M. |last4=Blondel |first4=V. D. |year=2013 |title=Unique in the Crowd: The privacy bounds of human mobility |url=https://www.nature.com/articles/srep01376 |journal=Scientific Reports |volume=3 |pages=1376 |doi=10.1038/srep01376 |access-date=2025-08-23}}</ref> Multiple peer-reviewed studies from 2018-2024 demonstrate that "anonymized" vehicle location data can be re-identified with high accuracy. A 2022 study showed researchers could re-identify drivers from raw vehicle network data with 97% accuracy by exploiting interdependencies in sensor measurements.<ref>{{cite web|url=https://www.sciencedirect.com/science/article/abs/pii/S0167404822002139|title=Privacy-preserving vehicle trajectory matching|website=ScienceDirect|date=2022|access-date=2025-10-05}}</ref> Research published in the Journal of Computer Science and Technology (2022) found that even 3-4 location points can uniquely identify individuals.<ref>{{cite journal|url=https://link.springer.com/article/10.1007/s11390-022-2409-x|title=Experiments and Analyses of Anonymization Mechanisms for Trajectory Data Publishing|journal=Journal of Computer Science and Technology|date=2022|access-date=2025-10-05}}</ref>
@@ Line 102: / Line 102: @@
 In Illinois, a Flock representative allegedly threatened a Department of Transportation official with police pressure when questioned about permit applications.<ref>{{Cite web |last=Uprise RI Staff |date=2024-10-23 |title=As Flock Surveillance Cameras Proliferate in Rhode Island, Lawsuit Challenges Their Legality |url=https://upriseri.com/as-flock-surveillance-cameras-proliferate-in-rhode-island-lawsuit-challenges-their-legality/ |url-status=live |access-date=2025-08-25 |website=UPRISE RI}}</ref>
+====='''In Evanston:'''=====
+Flock was ordered to remove 18 stationary cameras. The city put the contract with Flock on a 30-day termination notice on August 26. Flock Initially appeared to comply, removing 15 of the cameras by September 8. Later, <u>Flock was caught reinstalling all of them by the following Tuesday without authorization from the city.</u>
+The city of Evanston responded with a cease-and-desist order for Flock to remove the new and unauthorized camera equipment.
+Because Flock reinstalled the cameras without permission, <u>Evanston was forced to cover the cameras with tape and bags to block them from potentially logging vehicle data.</u>
+This situation is particularly absurd, as the city was left with little choice but to physically shield the cameras due to Flock’s unauthorized action, and outright noncompliance with the wishes of the city.
+Flock appears apprehensive to comply and remove the hardware, with a representative of flock stating:
+“We continue to be optimistic that we will have the opportunity to have a constructive dialogue to address the City’s concerns and resume our successful partnership making Evanston safer”
+While it is clear that flock wishes to continue their contract, this does not address the wishes of the city. In this regard, A Flock spokesperson stated: “the cameras are not active and will be uninstalled”
+This raises concerns, as <u>it is unclear why flock chose to incur costs to install</u> <u>unauthorized inactive cameras after the initial request for removal.</u> Additionally, it is not possible, or remains extremely difficult and costly for the city to independently verify that the ‘inactive’ cameras are truly disabled.
+It remains unclear if there is concrete financial incentive for flock to partake in these possibly illegal or otherwise unethical actions.
+More physical cameras are ostensibly capable of creating more data, which is collected in aggregate by flock, and selectively shared with other clients. Additional data could create a clear value proposition for certain flock services, and thus effects the desirability of the services provided by flock, and potentially, the price which flock can expect to charge for their services.<ref>{{Cite web |last=Harrison |first=Alex |date=September 25th, 2025 |title=City covers Flock cameras while waiting for removal |url=https://evanstonroundtable.com/2025/09/25/city-covers-up-flock-cameras-while-waiting-for-removal/ |url-status=live |archive-url=https://archive.ph/PD1qe |archive-date=2025-10-10 |access-date=2025-10-10 |website=Evanston Roundtable}}</ref>
 ===City rejections and terminations===
@@ Line 137: / Line 158: @@
 ===State audit findings===
-California State Auditor's February 2020 investigation found Los Angeles Police Department, with a 320 million image database, had no ALPR-specific policy at all. The audit found 96% of agencies claim to have policies, but most are incomplete. Data retention periods varied wildly with no justification. LAPD maintained a minimum five-year retention period, yet couldn't demonstrate that images stored for years had investigative value. The audit found that 99.9% of the 320 million images Los Angeles stores are for vehicles that were not on a hot list when the image was made.<ref>{{cite web|url=https://information.auditor.ca.gov/reports/2019-118/summary.html|title=Automated License Plate Readers|website=California State Auditor|date=2020-02|access-date=2025-10-05}}</ref>
+California State Auditor's February 2020 investigation found Los Angeles Police Department, with a 320 million image database, had no ALPR-specific policy at all. The audit found 96% of agencies claim to have policies, but most are incomplete. Data retention periods varied wildly with no justification. LAPD maintained a minimum five-year retention period, yet couldn't demonstrate that images stored for years had investigative value. The audit found that 99.9% of the 320 million images Los Angeles stores are for vehicles that were not on a hot list when the image was made.<ref>{{cite web|url=https://information.auditor.ca.gov/reports/2019-118/summary.html|title=Automated License Plate Readers|website=California State Auditor|date=2020-02-13|access-date=2025-10-05}}</ref>
 New Jersey provides a contrasting model with mandatory annual audits of all 523 law enforcement agencies. The 2024 audit reported only two significant violations, both involving users who hadn't completed required training.<ref>{{cite web|url=https://nj.gov/njsp/ALPR/pdf/2024_Audit_Automated_License_Plate_Recognition_(ALPR)_Data_Collected_Utilized_NJ_Law_Enforcement_Agencies.pdf|title=2024 Audit of Automated License Plate Recognition (ALPR) Data|website=New Jersey State Police|date=2024|access-date=2025-10-05}}</ref>
@@ Line 144: / Line 165: @@
 ===Cost-benefit analysis===
-Arizona Department of Transportation's 2008 study of generic ALPR technology (predating Flock Safety by nine years) estimated $9.98 million for a hypothetical statewide ALPR system. The projected benefit-to-cost ratio of 9.6:1 came entirely from registration and insurance compliance, not crime reduction.<ref>{{cite web|url=https://apps.azdot.gov/files/ADOTLibrary/publications/project_reports/PDF/AZ637.pdf|title=Automated License Plate Recognition Technology Implementation Report|website=Arizona Department of Transportation|date=2008-06|access-date=2025-10-05}}</ref>
+Arizona Department of Transportation's 2008 study of generic ALPR technology (predating Flock Safety by nine years) estimated $9.98 million for a hypothetical statewide ALPR system. The projected benefit-to-cost ratio of 9.6:1 came entirely from registration and insurance compliance, not crime reduction.<ref>{{cite web|url=https://apps.azdot.gov/files/ADOTLibrary/publications/project_reports/PDF/AZ637.pdf|title=Automated License Plate Recognition Technology Implementation Report|website=Arizona Department of Transportation|date=2008-06-01|access-date=2025-10-05}}</ref>
-Colorado's Office of Research and Statistics reported that while ALPR systems are expanding, independent academic research contradicts vendor claims. A 2011 George Mason University study concluded ALPRs "do not achieve a prevention or deterrent effect" on crime.<ref>{{cite web|url=https://cdpsdocs.state.co.us/ors/Docs/Briefs/2024-05_InDetail-ALPR.pdf|title=Automated License Plate Readers (In Detail)|website=Colorado Division of Criminal Justice|date=2024-05|access-date=2025-10-05}}</ref>
+Colorado's Office of Research and Statistics reported that while ALPR systems are expanding, independent academic research contradicts vendor claims. A 2011 George Mason University study concluded ALPRs "do not achieve a prevention or deterrent effect" on crime.<ref>{{cite web|url=https://cdpsdocs.state.co.us/ors/Docs/Briefs/2024-05_InDetail-ALPR.pdf|title=Automated License Plate Readers (In Detail)|website=Colorado Division of Criminal Justice|date=2024-05-01|access-date=2025-10-05}}</ref>
 Oakland Police Department reported 182 arrests from ALPR in the first year, representing 1.4% of homicides, robberies, burglaries, and firearm assaults. The Northern California Regional Intelligence Center states approximately 1-2 vehicles out of 1,000 initiate alerts—a hit rate of just 0.1-0.2%.<ref>{{cite web|url=https://ncric.ca.gov/wp-content/uploads/2021/10/California-Law-Enforcement-ALPR-FAQ.pdf|title=California Law Enforcement ALPR FAQ|website=NCRIC|date=2021|access-date=2025-10-05}}</ref>
@@ Line 153: / Line 174: @@
 The locations of many Flock Cameras have been mapped by the OpenStreetMap project.  A viewer of the locations of these cameras is located here: https://deflock.me/map
-==References==
+==== Cease and desist to Deflock.me ====
-{{reflist}}
+DeFlock.me is a website allowing users to log and view the locations of ALPRs, such as Flock products. On January 30th, 2025, Flock sent a cease and desist notice to the owner of DeFlock demanding the name of the website be changed to exclude the company's brand name. The letter also stated that "the Website also implies that various license plate readers are vulnerable to security hacks. . ." which Flock alleged ". . . provides a false impression about the security of Flock Products." <ref>''[https://www.eff.org/files/2025/02/26/2025_01_31_deflock_cd_ex-3.pdf “DEFLOCK”- Cease and Desist]. Institute for Justice. 2025-01-30. Retrieved 2025-10-17.''</ref>
 ==External links==
@@ Line 163: / Line 184: @@
 *[https://wiki.alprwatch.org/index.php/Main_Page ALPR Watch Wiki]
-[[Category:Surveillance]]
+==References==
-[[Category:Privacy violations]]
+{{reflist}}
-[[Category:Law enforcement technology]]
+[[Category:Flock License Plate Readers]]
-[[Category:Fourth Amendment issues]]
-[[Category:Data security]]