Issue
Mass Surveillance & Tracking
About this issue
Sources
- [1] Congressional Research Serviceprimary
“ALPRs are camera systems that capture the license plate data of passing vehicles, along with related information.”
- [2] Congressional Research Serviceprimary
“(ALPR technology can also detect additional, related information, including vehicle type and color, global positioning system [GPS] location data, and date and time.)”
- [3] Congressional Research Serviceprimary
“After they capture and catalog license plate information, ALPR systems can compare these data against various databases, including what are known as hot lists, which contain license plates linked to vehicles of interest.”
- [4] Congressional Research Serviceprimary
“nearly 90% of sheriffs' offices with 500 or more sworn deputies reported using the technology, and of police departments serving over 1 million residents, 100% used ALPRs.”
- [5] Congressional Research Serviceprimary
“network of LPR equipment owned by DEA, other federal agencies, and state, local, and tribal police departments. All law enforcement agencies involved have Memorandums of Understanding with DEA detailing the parameters for use and sharing of the LPR information.”
- [6] Congressional Research Serviceprimary
“No specific federal legislative framework exists that governs federal law enforcement use of ALPRs, though there are federal laws and policies broadly governing law enforcement investigations and intelligence gathering.”
- [24] USAspending.govprimary
“OBIM IS PROCURING THE MAINTENANCE FOR THE DHS PERPETUAL LICENSE FOR THE NEC AIM-XM AND AIM-ES PRODUCTS FOR UNLIMITED FACIAL QUANTITIES, ON UNLIMITED HARDWARE PLATFORMS, AND AT UNLIMITED LOCATIONS.”
- [25] USAspending.govprimary
“ENTERPRISE FACIAL LICENSE O&M”
- [26] Federal Trade Commissionprimary
“Rite Aid's facial recognition technology was more likely to generate false positives in stores located in plurality-Black and Asian communities than in plurality-White communities”
- [27] Federal Trade Commissionprimary
“The system generated thousands of false-positive matches, the FTC says.”
- [28] USAspending.govprimary
“THE UNITED STATES PARK POLICE (USPP) HAS A REQUIREMENT TO ACQUIRE AND DEPLOY A FLOCK SAFETY AUTOMATED LICENSE PLATE READER (ALPR) SYSTEM IN THE GREATER WASHINGTON, DC METROPOLITAN AREA (WMA).”
- [29] USAspending.govprimary
“LPR CAMERA SYSTEMS FOR DEAS NATIONAL LICENSE PLATE READER PROGRAM EXPANSION AND IMPROVEMENTS”
- [30] USAspending.govprimary
“REPAIR, PREVENTATIVE, REMEDIAL, AND ACCIDENTAL MAINTENANCE FOR CBP LICENSE PLATE READERS AND VEHICLE COUNTERS AT US PORTS OF ENTRY”
- [31] USAspending.govprimary
“IGF::CL::IGF ACCESS TO LICENSE PLATE READER DATABASE”
- [32] USAspending.govprimary
“CLEAR INVESTIGATIONS ADVANCED PLATFORM: REAL-TIME INCARCERATION AND ARREST, CLEAR ONLINE ALERTS, LEARN VIA VIGILANT SOLUTIONS FOR UP TO 750 USERS AND LPR VIA CLEAR FOR ALL CLEAR AUTHORIZED USERS.”
- [33] USAspending.govprimary
“SHADOWDRAGON HORIZON/SOCIALNET UNLIMITED QUERY, PER USER LICENSES (BASE YEAR) POP: 08/28/2023-08/27/2024”
- [34] ShadowDragonunknown
“Automate OSINT data collection and analysis across billions of public sources to dramatically reduce investigation time.”
- [35] Babel Streetunknown
“Unmask aliases and resolve identities across systems and languages to review the full picture.”
- [36] Dataminrunknown
“earliest warnings on high impact events and critical information far in advance of other sources.”
- [7] U.S. Department of Homeland Securityprimary
“This computer-readable format, also known as “a read,” contains some or all of the following information: (1) license plate number; (2) digital image of the license plate as well as the vehicle's make and model; (3) state of registration; (4) camera identification (i.e., camera owner and type); (5) Global Positioning Satellites coordinates or other location information recorded at the time the information was captured; and (6) date and time of observation.”
- [8] U.S. Department of Homeland Securityprimary
“The vendor compiles LPR records from at least 25 states and 24 of the top 30 most populous metropolitan statistical areas within the United States, to the extent that collection of LPR data is authorized by law in those jurisdictions.”
- [9] U.S. Department of Homeland Securityprimary
“The Contract for the commercial LPR service requires the vendor to provide an “alert list” feature that enables users to save license plate numbers to be automatically queried against new records as they are loaded into the vendor's LPR data service.”
- [10] U.S. Department of Homeland Securityprimary
“Records in EARM are maintained for 75 years.”
- [11] California State Auditorprimary
“All four agencies have accumulated a large number of images in their ALPR systems, yet most of the images do not relate to their criminal investigations—99.9 percent of the 320 million images Los Angeles stores are for vehicles that were not on a hot list when the image was made.”
- [12] California State Auditorprimary
“Furthermore, three of the four agencies have shared their ALPR images widely, without considering whether the entities receiving them have a right to and need for the images.”
- [13] California State Auditorprimary
“The statewide survey of law enforcement agencies we conducted found that 70 percent operate or plan to operate an ALPR system, and this raises concerns that these agencies may share the deficiencies we identified at the four agencies we reviewed.”
- [14] U.S. Department of Homeland Securityprimary
“Cell-site simulators permit the tracking of mobile telephones in the course of criminal investigations.”
- [15] U.S. Department of Homeland Securityprimary
“By transmitting as a cell tower, cell-site simulators acquire the identifying information from cellular devices.”
- [16] U.S. Department of Homeland Securityprimary
“During a target development mission, all mobile identifiers will be captured and processed afterward to identify the target number. Non-target data in the area will be deleted within 24 hours.”
- [17] USAspending.govprimary
“THE PURPOSE OF THIS DELIVERY ORDER IS TO PURCHASE VENNTEL GEOGRAPHIC MARKETING DATA AND SERVICE SUPPORT LICENSE SUBSCRIPTIONS FOR THE DHS SCIENCE AND TECHNOLOGY (S&T) DIRECTORATE, HOMELAND SECURITY ADVANCED RESEARCH PROJECTS AGENCY (HSARPA), DATA ANALYTICS ENGINE (DA-E).”
- [18] USAspending.govprimary
“THIS REQUIREMENT IS FOR 25 BABEL STREET INSIGHTS LICENSES, LOCATION AND IDENTITY APPLICATION PROGRAMMING INTERFACES FOR USE IN ICE CASE MANAGEMENT AND ANALYTICS SYSTEMS.”
- [19] USAspending.govprimary
“IGF::OT::IGF - 2018 TOG FOG DATA LES (LAW ENFORCEMENT SENSITIVE) $60K”
- [20] USAspending.govprimary
“THIS AWARD PROVIDES ICE HOMELAND SECURITY INVESTIGATIONS WITH PENLINKS COBWEB AND TANGLES LICENSES TO GATHER OPEN-SOURCE INFORMATION FOR HSI CRIMINAL INVESTIGATIONS.”
- [21] USAspending.govprimary
“THIS AWARD PROCURES FACIAL RECOGNITION SOFTWARE, WHICH SUPPORTS HOMELAND SECURITY INVESTIGATIONS WITH CAPABILITIES OF IDENTIFYING VICTIMS AND OFFENDERS IN CHILD SEXUAL EXPLOITATION CASES AND ASSAULTS AGAINST LAW ENFORCEMENT OFFICERS.”
- [22] USAspending.govprimary
“THE FR SYSTEM WITH THE LATEST STATE-OF-THE-ART FR MATCHING CAPABILITY, INTEGRATE THE FR CAPABILITY WITH THE DOS VISA AND PASSPORT ISSUANCE PROCESSES, TRANSITION FROM THE LEGACY FR SYSTEM INTERFACE, AND PROVIDE OPERATIONAL SUPPORT FOR THE NEW FR SYSTEM.”
- [23] USAspending.govprimary
“IDEMIA BIOMETRIC ALGORITHM MAINTENANCE FOR NGI”
Claims checked against the record
What companies and agencies say about this technology, set beside what the public record shows. The claim and the check are separate records with separate sources — a claim appearing here is not an endorsement of it, and a verdict is an assessment, open to dispute.
The claim
Flock Safety· Company website
“Using patented Vehicle Fingerprint™ technology, the camera captures the vehicle make, color, type, license plate, state of the license plate, missing plate, covered plate, paper plate, and over 20 unique vehicle details like roof racks and bumper stickers.”
Flock Safety's own marketing page describing what its cameras record about each passing vehicle.
Where this was said
- Flock Safetyunknown
“Using patented Vehicle FingerprintTM technology, the camera captures the vehicle make, color, type, license plate, state of the license plate, missing plate, covered plate, paper plate, and over 20 unique vehicle details like roof racks and bumper stickers.”
What the record shows
Partially supported
That a plate reader captures attributes beyond the plate is independently established: the Congressional Research Service records that the technology detects vehicle type and colour, and an ICE Privacy Impact Assessment describes a stored read as including the vehicle's make and model. Capture of the finer attributes Flock advertises — roof racks, bumper stickers, the 'over 20 unique vehicle details' — rests on Flock's own marketing alone; no government audit, court record or independent test confirming that list was located. Note also that Flock's current product pages have largely replaced the 'Vehicle Fingerprint' name with 'Vehicle Signature', so the wording above is from Flock's blog rather than its live product page.
Evidence
- Congressional Research Serviceprimary
“(ALPR technology can also detect additional, related information, including vehicle type and color, global positioning system [GPS] location data, and date and time.)”
- U.S. Department of Homeland Securityprimary
“(2) digital image of the license plate as well as the vehicle's make and model;”
Reviewer confidence 70%
- Flock Safetyunknown
The claim
Flock Safety· Company website
“Flock FreeForm expands on Vehicle Signature by going beyond preset filters, enabling searches based on everyday language, such as 'white sports car with a racing stripe' or 'red pickup truck with a dog in the bed.'”
Flock Safety's current product page, describing a natural-language search layer over its stored vehicle records.
Where this was said
- Flock Safetyunknown
“Flock FreeForm expands on Vehicle Signature by going beyond preset filters, enabling searches based on everyday language, such as 'white sports car with a racing stripe' or 'red pickup truck with a dog in the bed.'”
What the record shows
Not verifiable from public records
This is a product-capability claim on the vendor's own site. No independent evaluation, government audit or court record describing FreeForm's behaviour or accuracy was located, so it can be recorded as something Flock says but cannot be confirmed or refuted from public records. It is included because the claimed capability — free-text description search across stored vehicle records, including an occupant detail such as a dog in the truck bed — bears directly on what the stored data can be used to find.
Reviewer confidence 50%
- Flock Safetyunknown
The claim
Flock Safety· License Plate Reader Policy
“LPR data is hard deleted on a rolling 30-day basis by default; this may be increased or decreased on a case-by-case basis if a different schedule is required by a customer's law or policy.”
Flock Safety's published LPR policy, stating its default retention period.
Where this was said
- Flock Safetyunknown
“LPR data is hard deleted on a rolling 30-day basis by default; this may be increased or decreased on a case-by-case basis if a different schedule is required by a customer's law or policy.”
What the record shows
Supported by the record
Two independent records confirm the 30-day period in live deployments. The Court of Appeals of Virginia, describing Norfolk's system, found the images 'are saved in a searchable website database for 30 days'. The Village of Wilmette, Illinois states on its own municipal page that 'All ALPR data is only retained for 30 days.' Both are records outside Flock's control and both match the stated default. The claim is about the default only — Flock's own wording allows the period to be lengthened per customer — so this verdict does not establish that 30 days applies everywhere.
Evidence
- Court of Appeals of Virginiaprimary
“The images are saved in a searchable website database for 30 days and may be accessed by a Norfolk police officer to obtain information pertinent to a crime being investigated.”
- Village of Wilmette, Illinoisprimary
“All ALPR data is only retained for 30 days.”
Reviewer confidence 85%
- Flock Safetyunknown
The claim
Flock Safety· License Plate Reader Policy
“Customers choose whether to share LPR data with other customers in accordance with their laws and policies.”
Flock Safety's published policy on how data sharing between agencies is governed.
Where this was said
- Flock Safetyunknown
“Customers choose whether to share LPR data with other customers in accordance with their laws and policies.”
- Flock Safetyunknown
“to use the service, each Flock Safety partner agency must opt in to either the local or national sharing feature.”
What the record shows
Partially supported
The opt-in mechanism exists as described, but the qualifier 'in accordance with their laws and policies' describes an outcome the vendor does not enforce, and the public record shows it failing. The California Attorney General sued the City of El Cajon in October 2025 over sharing licence plate data with out-of-state agencies contrary to California law. Separately, the California State Auditor found that three of four audited agencies 'have shared their ALPR images widely, without considering whether the entities receiving them have a right to and need for the images', and that none had a policy meeting all legally mandated requirements. Placing the compliance decision with the customer is therefore accurate as a description of the mechanism and unreliable as an assurance of the result.
Evidence
- California Attorney Generalprimary
“ALPR systems are generally operated through third-party vendors that deploy cameras and collect and store ALPR data.”
- California State Auditorprimary
“Fresno and Marin have each arranged to share their ALPR images with hundreds of entities and Sacramento with over a thousand entities across the United States.”
- California State Auditorprimary
“None of the agencies have an ALPR usage and privacy policy that implements all the legally mandated—since 2016—requirements.”
Reviewer confidence 75%
- Flock Safetyunknown
The claim
Vigilant Solutions (Motorola Solutions)· LPR Usage and Privacy Policy
“The company retains LPR data as long as it has commercial value.”
Vigilant Solutions' own usage and privacy policy, describing retention of the plate data it collects commercially — as distinct from data collected by a police agency.
Where this was said
- Vigilant Solutionsunknown
“The company retains LPR data as long as it has commercial value.”
- Vigilant Solutionsunknown
“The company reserves the right to revise this policy at any point in the future and such changes will be retroactively applicable to data collected prior to any revision of this policy.”
What the record shows
Supported by the record
The statement is the company's own published policy and is quoted verbatim, so it is established as the stated practice. It is recorded here because it stands in sharp contrast to the 30-day default published by Flock Safety: commercially collected plate data is held on an open-ended commercial-value test rather than a fixed period. The same policy states that revisions apply retroactively to data already collected, meaning the retention terms attached to a past observation can change afterwards. That a federal agency purchases from this commercial pool is independently documented: a Department of State contract covers 'LEARN via Vigilant Solutions for up to 750 users'.
Evidence
- USAspending.govprimary
“CLEAR INVESTIGATIONS ADVANCED PLATFORM: REAL-TIME INCARCERATION AND ARREST, CLEAR ONLINE ALERTS, LEARN VIA VIGILANT SOLUTIONS FOR UP TO 750 USERS AND LPR VIA CLEAR FOR ALL CLEAR AUTHORIZED USERS.”
Reviewer confidence 80%
- Vigilant Solutionsunknown
The claim
Motorola Solutions· Company website
“Narrow results by make, model, color, accessories and even unique vehicle features like damage or bumper stickers.”
Motorola Solutions' licence plate recognition product page, describing attribute search beyond the plate number.
Where this was said
- Motorola Solutionsunknown
“Narrow results by make, model, color, accessories and even unique vehicle features like damage or bumper stickers.”
- Motorola Solutionsunknown
“Transform license plates into leads with patented, powerful vehicle location analytics and access to billions of detections beyond your own.”
What the record shows
Partially supported
Attribute capture beyond the plate is independently established in general terms by the Congressional Research Service (vehicle type and colour) and by an ICE Privacy Impact Assessment (make and model). Detection of damage and bumper stickers specifically, and the scale implied by 'access to billions of detections beyond your own', rest on the company's own marketing; no independent test or audit of either was located.
Evidence
- Congressional Research Serviceprimary
“(ALPR technology can also detect additional, related information, including vehicle type and color, global positioning system [GPS] location data, and date and time.)”
Reviewer confidence 65%
- Motorola Solutionsunknown
The claim
Leonardo (ELSAG)· SignalTrace product page
“SignalTrace creates an electronic fingerprint for groups of the signals that the system determines are frequently emitted together.”
Leonardo's own product page for SignalTrace, an add-on to its ELSAG plate-reader line that identifies vehicles by the radio signals emitted from devices inside them rather than by the licence plate.
Where this was said
- Leonardo US Cyber and Security Solutionsunknown
“SignalTrace creates an electronic fingerprint for groups of the signals that the system determines are frequently emitted together.”
- Leonardo US Cyber and Security Solutionsunknown
“Electronic devices such as fitness trackers, smartwatches, RFID tags and mobile phones emit signals into the air.”
What the record shows
Not verifiable from public records
Recorded because of what it describes rather than because it can be checked: a plate-reader vendor marketing identification of a vehicle by the combined radio signatures of the personal devices travelling in it, which does not require a licence plate at all. The capability claim comes solely from Leonardo's own product page; no procurement record, audit or court document describing a SignalTrace deployment at a named agency was located, so it can be neither confirmed nor refuted from public records.
Reviewer confidence 50%
- Leonardo US Cyber and Security Solutionsunknown
The claim
Axon Enterprise· Company announcement, 23 October 2019
“We have not, and will not ever sell public safety data.”
Axon's announcement of its in-car licence plate reader, committing not to sell data collected through it.
Where this was said
- Axonunknown
“We have not, and will not ever sell public safety data.”
What the record shows
Not verifiable from public records
A forward-looking commitment about future conduct cannot be established or refuted from the public record. No evidence contradicting it was located, and none confirming ongoing compliance either. It is recorded so the commitment is on the record and can be checked against later conduct.
Reviewer confidence 50%
- Axonunknown
The claim
Rekor Systems· Product documentation
“Rekor's vehicle recognition utilizes artificial intelligence and machine learning to identify over 2000 unique vehicle characteristics and capture the complete vehicle signature.”
Rekor's own developer documentation, which publishes the most detailed attribute catalogue located for any vendor in this issue.
Where this was said
- Rekor Systemsunknown
“Rekor's vehicle recognition utilizes artificial intelligence and machine learning to identify over 2000 unique vehicle characteristics and capture the complete vehicle signature.”
- Rekor Systemsunknown
“Using custom, shared, and NCIC hotlists; receive alerts on vehicles of interest the moment they are detected.”
What the record shows
Not verifiable from public records
The attribute catalogue is published by the vendor in its own technical documentation, which establishes what Rekor says its software identifies but not its accuracy in the field. No independent evaluation was located. Rekor's related claim to hold a patent covering point-of-collection anonymisation was not checked against the patent record and is not asserted here.
Reviewer confidence 50%
- Rekor Systemsunknown
The claim
International Association of Chiefs of Police / U.S. Department of Justice· NIJ-funded operational guidance, NCJ 239604
“A total of 40% of respondents retain ALPR data for six months or less (n=16). Five respondents (13%) indicated they retain ALPR data indefinitely, while two indicated that retention is based on the storage capacity of the equipment installed.”
Federally funded guidance surveying how long police agencies actually keep plate reader data — included as the counterpoint to vendors' stated default retention periods.
Where this was said
- U.S. Department of Justice, National Institute of Justiceprimary
“A total of 40% of respondents retain ALPR data for six months or less (n=16). Five respondents (13%) indicated they retain ALPR data indefinitely, while two indicated that retention is based on the storage capacity of the equipment installed.”
What the record shows
Supported by the record
This is a federally published survey finding, primary for what agencies reported. It matters here because it shows retention is set by each agency rather than by any vendor default: 13% of responding agencies reported keeping plate data indefinitely, and two set retention by whatever their hardware could store. A vendor's stated default period therefore describes the vendor's configuration, not the life of the data. The California State Auditor reached a compatible conclusion, finding 99.9 percent of the 320 million images held by one agency related to vehicles not on any hot list.
Evidence
- California State Auditorprimary
“99.9 percent of the 320 million images Los Angeles stores are for vehicles that were not on a hot list when the image was made.”
Reviewer confidence 80%
- U.S. Department of Justice, National Institute of Justiceprimary
The claim
Flock Safety· Company website — Privacy & Trust FAQ
“No. Flock cameras do not use facial recognition technology. Searches are based on vehicle characteristics, and not identity. Flock cannot recognize, identify, or track individuals. It is purpose built for specific public safety investigations, not to monitor people.”
Answer to 'Do Flock cameras use facial recognition?' on Flock's Privacy & Trust hub. Retrieved 2026-08-02.
Where this was said
- Flock Safetyunknown
“No. Flock cameras do not use facial recognition technology. Searches are based on vehicle characteristics, and not identity. Flock cannot recognize, identify, or track individuals.”
What the record shows
Partially supported
The narrow half holds: no record was found contradicting the statement that Flock's plate readers do not run facial recognition. The broader half — that Flock 'cannot recognize, identify, or track individuals' — is undercut by Flock's own product documentation. Its FreeForm page advertises plain-language search for people by physical description and returns 'person-related video search results'. That is recognition of individuals by attribute, even though it is not biometric face matching. The distinction Flock draws is real; the sentence as written overstates it.
Evidence
- Flock Safetyunknown
“Type what you're looking for, such as "man in blue shirt and cowboy hat" or "dark SUV with rear bumper damage." Then review relevant visual matches across enabled video feeds, with vehicle results also available from LPR systems.”
- Flock Safetyunknown
“Review person-related video search results from enabled video feeds alongside vehicle results from video and LPR systems in one interface.”
Reviewer confidence 80%
- Flock Safetyunknown
The claim
Flock Safety· Company website — Myths vs Facts page
“Flock clients control their data; no secret federal agency access or "backdoors" exist.”
Answer to 'Do Federal Agencies Have Access?' on Flock's Myths vs Facts page. Retrieved 2026-08-02.
Where this was said
- Flock Safetyunknown
“Flock clients control their data; no secret federal agency access or "backdoors" exist.”
What the record shows
Refuted by the record
The Illinois Secretary of State's audit found the opposite as a matter of fact for the period audited: U.S. Customs and Border Protection obtained access to Illinois plate-reader data, Flock lacked adequate data-sharing safeguards, and Flock was running a CBP pilot that its own leadership did not know about. A company whose leadership is unaware of a federal agency's access to customer data cannot at the same time assure customers that no undisclosed federal access exists. Flock's own subsequent disclosure corroborates the pilot, dating it 9 May to 24 August 2025. This refutes the claim as a description of how the system operated through August 2025; whether it holds today, after the remediation Flock describes, is not established either way by the records reviewed.
Evidence
- Illinois Secretary of Stateprimary
“During a recent audit, Giannoulias' office found that Flock Safety, which operates the largest automated license plate reader (ALPR) system in the nation, allowed U.S. Customs and Border Protection to access Illinois license plate cameras on Illinois roads and surveil drivers.”
- Illinois Secretary of Stateprimary
“the office discovered late last week that Flock did not have proper safeguards in place for data sharing, which was compounded by the fact that the company was running a pilot program with U.S. Customs and Border Protection, which Flock leadership was unaware of.”
- Flock Safetyunknown
“CBP Pilot: May 9, 2025 to August 24, 2025”
Reviewer confidence 80%
- Flock Safetyunknown
The claim
Flock Safety· Company blog — official position paper
“No. Flock does not work with U.S. Immigration and Customs Enforcement (ICE). Communities control federal data access. ICE does not have direct access to Flock cameras, systems, or data, unless the agencies that control their data expressly and deliberately allow it.”
Opening answer of Flock's blog post 'Does Flock Share Data With ICE?', published 6 January 2026 and updated 24 July 2026.
Where this was said
- Flock Safetyunknown
“No. Flock does not work with U.S. Immigration and Customs Enforcement (ICE). Communities control federal data access.”
What the record shows
Partially supported
The narrowest assertion — no ICE contract and no direct ICE login — is not contradicted by any record retrieved, and Flock's own published pilot timeline lists no ICE pilot. But the framing that federal access happens only when local agencies 'expressly and deliberately allow it' is undercut by the Illinois audit, which found CBP access that customer agencies had not authorised. Flock's own post also concedes pilots with CBP, HSI, NCIS, FBI and ATF, and describes disabling National Lookup for all California agencies in March 2025 — a Flock-level control action, not a customer one. Accurate on its literal terms; misleading as a general assurance about federal access.
Evidence
- Flock Safetyunknown
“In March 2025, Flock disabled National Lookup for all California agencies.”
- Illinois Secretary of Stateprimary
“This sharing of license plate data of motorists who drive on Illinois roads is a clear violation of the state law.”
- Flock Safetyunknown
“In August of 2025, Flock publicly announced it would no longer conduct pilot projects with federal agencies.”
Reviewer confidence 75%
- Flock Safetyunknown
The claim
Flock Safety· Company legal policy — LPR Usage and Privacy Policy
“LPR data gathered by the Flock Safety system on behalf of Flock Safety customers is owned by the customer.”
Under 'Restrictions on the Sale, Sharing, or Transfer of LPR Data' in Flock's binding LPR Policy, last updated 30 June 2026.
Where this was said
- Flock Safetyunknown
“LPR data gathered by the Flock Safety system on behalf of Flock Safety customers is owned by the customer.”
What the record shows
Partially supported
The ownership statement is a real contractual commitment, and the same policy binds Flock not to sell customer plate data. Two carve-outs in that same document qualify it: Flock reserves the right to disclose the data to government officials and third parties on its own good-faith judgement, and it retains a fraction of customer images for its own machine-learning development. Ownership in the sense of exclusive customer control over disposition is therefore not what the policy establishes. The Illinois audit further shows ownership did not translate into control in practice, since federal access occurred without the owning agencies' authorisation.
Evidence
- Flock Safetyunknown
“Flock Safety may access, use, preserve and/or disclose the LPR data to law enforcement authorities, government officials, and/or third parties, if legally required to do so or if Flock has a good faith belief that such access, use, preservation or disclosure is reasonably necessary to comply with a legal process, enforce the agreement between Flock and the customer, or detect, prevent or otherwise address security, privacy, fraud or technical issues.”
- Flock Safetyunknown
“Additionally, Flock uses a fraction of LPR images (less than one percent), which are stripped of all metadata and identifying information, solely for the purpose of improving Flock Services through machine learning.”
- Illinois Secretary of Stateprimary
“As a result, Giannoulias immediately ordered the company to shut off access to U.S. Customs and Border Protection.”
Reviewer confidence 75%
- Flock Safetyunknown
The claim
Flock Safety· Company website — Myths vs Facts page
“Flock gives customers full data ownership, auto-deleting data after 30 days, and prevents unauthorized sharing.”
Answer to 'How Is Data Protected?' on Flock's Myths vs Facts page. Retrieved 2026-08-02.
Where this was said
- Flock Safetyunknown
“Flock gives customers full data ownership, auto-deleting data after 30 days, and prevents unauthorized sharing.”
What the record shows
Partially supported
Flock's own binding policy makes 30 days a DEFAULT that can be lengthened or shortened per customer, so 'auto-deleting data after 30 days' states as a fixed property what the policy states as adjustable. The California State Auditor's review is the closest governmental test of whether retention defaults hold in practice, and found agencies retaining images for one to five years. That audit examined Los Angeles, Fresno, Sacramento County and Marin County programmes — NOT Flock deployments — so it establishes the general pattern rather than a Flock-specific one. The 'prevents unauthorized sharing' element is separately contradicted by the Illinois audit.
Evidence
- Flock Safetyunknown
“LPR data is hard deleted on a rolling 30-day basis by default; this may be increased or decreased on a case-by-case basis if a different schedule is required by a customer's law or policy.”
- California State Auditorprimary
“Fresno's policy is to retain ALPR images for one year; Sacramento's and Marin's policies specify two years.”
- Illinois Secretary of Stateprimary
“Flock did not have proper safeguards in place for data sharing”
Reviewer confidence 80%
- Flock Safetyunknown
The claim
Flock Safety· Company website — Myths vs Facts page
“Data sharing is controlled by the customer agency, not Flock.”
Answer to 'Who Can Access Data?' on Flock's Myths vs Facts page. Retrieved 2026-08-02.
Where this was said
- Flock Safetyunknown
“Data sharing is controlled by the customer agency, not Flock.”
What the record shows
Partially supported
Customer-level control is genuine in one direction: the California Attorney General sued the City of El Cajon over sharing its own plate data with agencies in more than two dozen states — an enforcement action against the agency, not the vendor, consistent with the agency holding the sharing decision. But control is not exclusive. The Illinois audit found federal access flowing from a Flock-run pilot rather than a customer decision, and Flock itself describes unilaterally disabling National Lookup for all California agencies. Both show Flock exercising sharing control independent of the customer.
Evidence
- California Office of the Attorney Generalprimary
“the City of El Cajon Police Department continues to share this data with numerous out-of-state law enforcement agencies throughout the country”
- Illinois Secretary of Stateprimary
“Flock did not have proper safeguards in place for data sharing”
- Flock Safetyunknown
“In March 2025, Flock disabled National Lookup for all California agencies.”
Reviewer confidence 75%
- Flock Safetyunknown
The claim
Flock Safety· Company website — Privacy & Trust FAQ
“No. Access is managed by your local agency and limited to authorized users with role-based permissions. Every search is tied to a specific user and recorded automatically. There is no open browsing and no anonymous searches. Access is controlled locally, not by Flock.”
Answer to 'Can anyone access my data at any time?' on Flock's Privacy & Trust hub. Retrieved 2026-08-02.
Where this was said
- Flock Safetyunknown
“Every search is tied to a specific user and recorded automatically. There is no open browsing and no anonymous searches.”
What the record shows
Partially supported
The logging mechanism demonstrably exists as described: Flock's policy specifies that every query stores username, date, time, stated purpose and search terms, and the Johnson County, Texas search became public precisely because those audit records were obtainable. What the record does not support is the inference drawn from it — that logging delivers accountability. The stated 'purpose of query' is free text supplied by the searching officer and is not validated, and the California State Auditor found agencies had failed to audit their systems at all, leaving them open to abuse. Logging creates a reviewable record; it does not by itself constrain who searches, or why.
Evidence
- California State Auditorprimary
“The agencies we reviewed have few safeguards for the creation of ALPR user accounts and have also failed to audit the use of their ALPR systems. Instead of ensuring that only authorized users access ALPR data for appropriate purposes, the agencies have left their systems open to abuse by neglecting to institute sufficient oversight.”
- Electronic Frontier Foundationadvocacy
“being searched for as a missing person, not as a suspect of a crime”
- Flock Safetyunknown
“All queries of the LPR system are stored for auditing purposes, including: Username Date Time Purpose of query License plate and other elements used to query the system”
Reviewer confidence 85%
- Flock Safetyunknown
The claim
Flock Safety· Company blog — announcement of company-run study
“Flock technology helps resolve roughly 700,000 crimes a year across law enforcement, businesses, neighborhoods, and schools: about 10% of reported U.S. crime.”
Flock blog post '10% of All Reported U.S. Crime Is Solved Using Flock', announcing a study Flock commissioned and staffed. Published February 2024, still live August 2026.
Where this was said
- Flock Safetyunknown
“Flock technology helps resolve roughly 700,000 crimes a year across law enforcement, businesses, neighborhoods, and schools: about 10% of reported U.S. crime.”
What the record shows
No supporting record found
The figure rests on a study Flock itself conducted, and one of the two academic co-authors who provided the independent oversight Flock cited has publicly disavowed the methodology, saying the underlying police data were too varied and incomplete for meaningful statistical analysis. No peer-reviewed replication exists. The strongest independent local test points the other way: Oak Park, Illinois's Citizens Police Oversight Committee found no evidence Flock cameras had played a meaningful role in any local investigation over three years. UNSUPPORTED rather than REFUTED — one village cannot disprove a national estimate, but no independent evidence establishes the 10% figure either.
Evidence
- 404 Mediaunknown
“the information that is collected by the police departments are too varied and incomplete for us to do any type of meaningful statistical analysis on them”
- Institute for Justiceunknown
“There is no evidence whatsoever that Flock Safety ALPRs have played a meaningful role in any Oak Park crime investigation since their installation in 2022,”
Reviewer confidence 80%
- Flock Safetyunknown
The claim
Flock Safety· Company blog — product marketing
“Flock's LPR is highly accurate, achieving over 99% capture in clear and rainy conditions and over 98% at dawn and dusk. It also delivers over 96% OCR accuracy and over 97% license plate state accuracy.”
Flock blog post 'Vehicle Recognition Software for Investigation-grade LPR'. Retrieved 2026-08-02.
Where this was said
- Flock Safetyunknown
“Flock's LPR is highly accurate, achieving over 99% capture in clear and rainy conditions and over 98% at dawn and dusk.”
What the record shows
Partially supported
Flock's own binding policy is more guarded than its marketing, warning that translation 'may be incomplete or inaccurate' and instructing users to confirm a read before acting on an alert — which sits awkwardly with treating the advertised percentages as operationally reliable. Public records from Roseville, California covering 1,427 alerts were reported to show 71% of alerts tied to stolen or felony vehicles resting on incorrect reads; Flock disputes that characterisation and attributes it to non-standard hardware and placement, and its published rates are scoped to optimal conditions, so the two figures are not strictly contradictory. Separately, the Institute for Justice documented at least 27 mistaken stops since 2018. Note the underlying Roseville records were not obtained directly; that figure rests on reporting.
Evidence
- Flock Safetyunknown
“Although infrequent, license plate translation may be incomplete or inaccurate. False positives are flagged to continuously improve the system. Users should confirm the computer translation prior to taking any action based on an LPR alert or search.”
- Gizmodounknown
“71% of cases that generated an alert related to vehicles that were stolen or used in a felony”
- Institute for Justiceunknown
“in nearly two-thirds of the cases IJ analyzed, officers did not realize their error until after they had drawn and pointed their guns at innocent people.”
Reviewer confidence 70%
- Flock Safetyunknown
The claim
Flock Safety· Company website — Myths vs Facts page
“Flock LPR captures public vehicle info at a single point in time, not continuous movement.”
Answer to 'Are You Being Tracked?' on Flock's Myths vs Facts page. Retrieved 2026-08-02.
Where this was said
- Flock Safetyunknown
“Flock LPR captures public vehicle info at a single point in time, not continuous movement.”
What the record shows
Refuted by the record
The statement is true of one camera and false of the system Flock sells. Flock's own product page advertises the 'Largest Fixed LPR Network' with billions of monthly reads, and its vehicle-recognition post puts the platform at more than 20 billion reads per month across 49 states with cross-agency sharing. A network that indexes a plate's appearances across tens of thousands of cameras and retains them for 30 days by default reconstructs movement over time — that is the stated investigative value of the product. Note this verdict rests on Flock's own product documentation; no court ruling holding that Flock's network specifically constitutes long-term tracking was located.
Evidence
- Flock Safetyunknown
“With billions of monthly plate reads, Flock connects communities, businesses and law enforcement in a shared network built to stop crime.”
- Flock Safetyunknown
“The platform processes 20B+ vehicle reads per month across 49 states, with thousands of law enforcement agencies participating in controlled sharing.”
- Flock Safetyunknown
“LPR data is hard deleted on a rolling 30-day basis by default”
Reviewer confidence 75%
- Flock Safetyunknown
The claim
Flock Safety· Dan Haley, Chief Legal Officer
“Today's ruling reinforces what courts nationwide have recognized: LPRs are lawful.”
Flock press statement responding to Commonwealth v. Church (Va. Ct. App., 14 October 2025). The same post asserts courts 'have uniformly upheld the constitutionality of LPRs'.
Where this was said
- Flock Safetyunknown
“Today's ruling reinforces what courts nationwide have recognized: LPRs are lawful.”
What the record shows
Partially supported
Flock accurately reports its win in the Virginia Court of Appeals, and that holding is consistent with a line of cases treating plate capture on public roads as outside the Fourth Amendment. But 'uniformly upheld' overstates a moving body of law. The Supreme Court's decision in Chatrie v. United States, issued 29 June 2026, held that acquiring a person's aggregated location data is a Fourth Amendment search and declined to extend the third-party doctrine to it. That reasoning is directed at cell-phone location data, NOT plate readers, but it bears squarely on the aggregation theory underlying challenges to networked plate data. The claim also concerns every court in the country, which could not be exhaustively verified.
Evidence
- Supreme Court of the United Statesprimary
“Held: Police officers conducted a Fourth Amendment search when they acquired Chatrie's location data from Google because an individual has a reasonable expectation of privacy in his cell-phone location information.”
- Supreme Court of the United Statesprimary
“Where the Fourth Amendment applies, it applies regardless of "the quality or quantity of information" the government obtains.”
Reviewer confidence 65%
- Flock Safetyunknown
The claim
Flock Safety· Company website — product page statistic
“52% Reduction in robbery in 1 year”
Headline statistic on Flock's licence plate reader product page, above a customer testimonial. No jurisdiction, baseline, time period or methodology is given. Retrieved 2026-08-02.
Where this was said
- Flock Safetyunknown
“52% Reduction in robbery in 1 year”
What the record shows
Not verifiable from public records
The figure is published without a named jurisdiction, date range, comparison group or source, so there is nothing specific to check it against. What can be said is that the rigorous evaluation literature does not support a general deterrent effect of this magnitude: a randomised controlled experiment by George Mason University, funded by the National Institute of Justice, found no crime reduction from plate-reader deployment, and characterised the overall evidence base as weak. That does not disprove a particular agency's result, which could reflect other simultaneous interventions — it means the claim as published cannot be evaluated.
Evidence
- Center for Evidence-Based Crime Policy, George Mason Universityunknown
“Our randomized controlled experiment mirrored the findings from the PERF experiments in that the use of LPR in autotheft hot spots does not appear to result in a reduction of crime generally or autotheft specifically, during the period of time measured.”
- Center for Evidence-Based Crime Policy, George Mason Universityunknown
“we also discovered this rapid adoption is occurring in a low-information environment; the evidence-base for the effectiveness and effects of LPR is weak.”
Reviewer confidence 60%
- Flock Safetyunknown
What the public thinks
An opinion poll measures what people believe. It is not evidence of what is true, and nothing here affects any promise rating. Figures are reported exactly as each source published them and are never averaged or combined across pollsters.
Government statistics
Bureau of Justice Statistics
Public experience of police misconduct during contacts
- Experienced any misconduct during most recent police contact (2022)1.2%
- Experienced any misconduct during most recent police contact (2020)1.1%
- Police behaviour seemed motivated by prejudice or bias (2022)1.1%
Subgroups (3)
- White (2022) — Any misconduct0.7%
- Black (2022) — Any misconduct4.2%
- Hispanic (2022) — Any misconduct1.4%
U.S. residents age 16+ who had contact with police · Supplement to the National Crime Victimization Survey, a nationally representative household survey · source
U.S. Census Bureau for NTIA
Online privacy and security concerns (2019 NTIA Internet Use Survey)
- Had significant concerns about online privacy and security risks (2019)73%
- Held back from some online activities as a result (2019)35%
- Concerned about data collection by online services (2019)26%
- Concerned about data collection by online services (2015)23%
U.S. households with at least one internet user · Biennial Computer and Internet Use Supplement to the Current Population Survey · source
U.S. Census Bureau for NTIA
Online privacy and security concerns (2015 NTIA Internet Use Survey)
- Named at least one concern about online privacy and security risks84%
- Named identity theft as a major concern63%
- Refrained from an online activity (transactions, purchases, posting, expressing opinions) due to privacy/security concerns45%
- Reported being affected by a breach, identity theft or similar in the prior 12 months19%
Subgroups (2)
- Affected by a breach in the prior year — Concerned about data collection or tracking by online services30%
- Not affected by a breach — Concerned about data collection or tracking by online services21%
U.S. households with at least one internet user · Computer and Internet Use Supplement to the Current Population Survey; open-ended concern question, multiple answers allowed · source
Private opinion polling — not evidence
Published by private polling organisations. Shown for context only; each is reported separately, with its question wording, so you can judge it yourself.
The Harris Poll
Automated licence-plate readers and AI cameras (VENDOR-COMMISSIONED)
- Say signs should be visible notifying people that plate readers are in use87%
- Agree businesses may use facial recognition to match against a list of known violent offenders79%
- Feel safer parking in lots that use automated plate readers78%
- Believe non-suspicious licence plate data should be deleted within 30 days67%
- Support AI video cameras in public places to identify suspicious behaviour before it occurs60%
U.S. adults 18+ · n=2,089 · ±2.7 · SPONSORSHIP CAVEAT: commissioned by a company that sells the surveillance cameras the survey asks about, and distributed as a corporate press release. Items are positively-worded agree/disagree statements, which typically inflates agreement. Included because it was the only readable national ALPR polling located; treat as sponsor-framed · 2026-03-23 – 2026-03-25 · source
Quinnipiac University Poll
Military use of AI in surveillance
“Do you support or oppose the military using AI in surveillance for security purposes?”
- Support45%
- Oppose44%
- Don't know / no answer11%
Subgroups (3)
- Gen Z (1997-2008) — Oppose58%
- Millennials (1981-1996) — Oppose49%
- Baby Boomers (1946-1964) — Support53%
U.S. adults · n=1,397 · ±3.3 · Probability-based random digit dialling with live interviewers. Concerns MILITARY surveillance, not domestic policing · 2026-03-19 – 2026-03-23 · source
YouGov
Government surveillance and privacy
- Acceptable to monitor the online activity of suspected terrorists84%
- Acceptable to monitor the online activity of politicians66%
- Say collecting Americans' phone and internet data is an unnecessary intrusion56%
- Say it is justified as a way to combat terrorism and crime24%
- Concerned surveillance powers could be used to target political opponents or suppress dissent71%
- Say government surveillance makes people less likely to express political views online47%
- Have chosen not to post or send something over monitoring concerns24%
- Say government surveillance has increased in the past 25 years78%
Subgroups (5)
- Democrats — Very concerned surveillance could target political opponents or suppress dissent55%
- Republicans — Very concerned surveillance could target political opponents or suppress dissent23%
- Democrats — Have self-censored online over monitoring concerns33%
- Republicans — Have self-censored online over monitoring concerns16%
- Republicans — Acceptable to monitor the online activity of journalists56%
U.S. adult citizens · n=1,131 · ±4 · NON-PROBABILITY: YouGov opt-in online panel, weighted to be representative of adult U.S. citizens. Verbatim topline wording is not printed in the article · 2025-05-30 – 2025-06-03 · source
Pew Research Center
Acceptability of law enforcement surveillance tools in investigations
- Obtain footage from residential cameras such as video doorbells — acceptable74%
- Use cellphone tower information to track where someone is — acceptable74%
- Require third parties to turn over private chats, messages or calls — acceptable55%
- Break the passcode on a user's phone — acceptable54%
Subgroups (6)
- Ages 18-29 — Use cellphone tower information to track someone — acceptable57%
- Ages 65+ — Use cellphone tower information to track someone — acceptable86%
- Black — Use cellphone tower information to track someone — acceptable66%
- White — Use cellphone tower information to track someone — acceptable77%
- Ages 18-29 — Break the passcode on a user's phone — acceptable35%
- Ages 65+ — Break the passcode on a user's phone — acceptable68%
U.S. adults · n=5,101 · ±1.7 · American Trends Panel Wave 127, split-half design. Percentages are the very/somewhat acceptable NET as published; verbatim topline wording is not printed in the report · 2023-05-15 – 2023-05-21 · source
Pew Research Center
Concern about government use of collected data
“How concerned are you about how the government is using the data it collects about you?”
- Very concerned (2023)31%
- Somewhat concerned (2023)39%
- Not too concerned (2023)24%
- Not at all concerned (2023)5%
- Very concerned (2019)25%
Subgroups (2)
- Republicans and GOP leaners (2023) — Very or somewhat concerned77%
- Republicans and GOP leaners (2019) — Very or somewhat concerned63%
U.S. adults · n=5,101 · ±1.7 · American Trends Panel Wave 127. Pew publishes the combined very/somewhat NET as 71% for 2023 and 64% for 2019 — do not re-derive from the rounded components · 2023-05-15 – 2023-05-21 · source
Monmouth University Polling Institute
Facial recognition in public places such as stadiums
“Facial recognition technology that can identify and monitor people who try to enter public places such as stadiums — is that a good idea or bad idea?”
- Good idea54%
- Bad idea41%
Subgroups (2)
- Have heard a lot about recent AI developments — Good idea45%
- Have heard nothing about recent AI developments — Good idea71%
U.S. adults · n=805 · ±5.7 · Telephone interviews with live interviewers; probability-based national random sample · 2023-01-26 – 2023-01-30 · source
Pew Research Center
Regulation of police facial recognition
“If the use of facial recognition technology by police becomes widespread, which is your greater concern?”
- Government will not go far enough regulating its use51%
- Government will go too far regulating its use47%
Subgroups (4)
- Rep/lean Rep — Government will go too far regulating its use59%
- Dem/lean Dem — Government will go too far regulating its use36%
- Dem/lean Dem — Government will not go far enough regulating its use62%
- Rep/lean Rep — Government will not go far enough regulating its use40%
U.S. adults · n=5,153 · American Trends Panel, Form 1 only · 2021-11-01 – 2021-11-07 · source
Pew Research Center
Facial recognition match as evidence for arrest
“If a facial recognition program said that someone was involved in a crime, should that be good enough evidence for police to arrest them, even if there was a small chance the program was wrong?”
- Should NOT be good enough evidence to arrest70%
- Should be good enough evidence to arrest27%
Subgroups (2)
- Say police facial recognition is a good idea for society — Should be good enough evidence to arrest40%
- Say police facial recognition is a bad idea for society — Should be good enough evidence to arrest15%
U.S. adults · n=5,153 · American Trends Panel, Form 1 only · 2021-11-01 – 2021-11-07 · source
Pew Research Center
Acceptable contexts for police facial recognition scanning
- Scanning people as they enter large events like concerts — acceptable63%
- Scanning people at public protests — acceptable61%
- Scanning people as they walk down the street — acceptable31%
- Scanning people as they walk down the street — NOT acceptable68%
U.S. adults · n=5,153 · American Trends Panel, Form 1 only. Verbatim topline wording is not printed in the report, so the item texts are reproduced from the published chart labels · 2021-11-01 – 2021-11-07 · source
Pew Research Center
Expected consequences of widespread police facial recognition
“If the use of facial recognition technology by police becomes widespread, do you think each of the following would happen? The police would…”
- Be able to track everyone's location at all times — definitely would happen27%
- Be able to track everyone's location at all times — probably would happen42%
- Monitor Black and Hispanic neighborhoods much more often — definitely would happen26%
- Monitor Black and Hispanic neighborhoods much more often — probably would happen40%
- Make more false arrests — definitely would happen16%
- Make more false arrests — probably would happen38%
- Solve crimes more quickly and efficiently — probably would happen55%
- Find more missing persons — probably would happen57%
Subgroups (5)
- Black adults — Definitely would monitor Black and Hispanic neighborhoods much more often48%
- Hispanic adults — Definitely would monitor Black and Hispanic neighborhoods much more often37%
- White adults — Definitely would monitor Black and Hispanic neighborhoods much more often18%
- Black adults — Definitely would make more false arrests28%
- White adults — Definitely would make more false arrests11%
U.S. adults · n=5,153 · American Trends Panel, Form 1 only; items randomised · 2021-11-01 – 2021-11-07 · source
Pew Research Center
Police facial recognition — good or bad idea for society
“Do you think the widespread use of facial recognition technology by police would be a…”
- Good idea for society46%
- Bad idea for society27%
- Not sure27%
U.S. adults · n=5,153 · American Trends Panel, Form 1 only. Pew's published ±1.6 applies to the full 10,260 wave, not this subsample, so no margin of error is shown · 2021-11-01 – 2021-11-07 · source
AP-NORC Center for Public Affairs Research
Surveillance cameras in public places
“Now you will read a list of specific policy measures about things happening in public places that some people have suggested the government should take on in response to threats against the United States.”
- Favor surveillance cameras in public places to watch for suspicious activity (NET)60%
- Neither favor nor oppose21%
- Oppose (NET)18%
- Oppose racial and ethnic profiling for tougher airport screening (NET)61%
Subgroups (3)
- Asian Americans — Oppose racial and ethnic profiling for tougher airport screening71%
- Black Americans — Oppose racial and ethnic profiling for tougher airport screening68%
- White Americans — Oppose racial and ethnic profiling for tougher airport screening59%
U.S. adults · n=1,729 · ±3.2 · AmeriSpeak probability-based panel. AP-NORC prints 2013 (72% favor) and 2011 (71%) trend rows, but those waves were telephone-only, a different mode · 2021-08-12 – 2021-08-16 · source
AP-NORC Center for Public Affairs Research
Warrantless government surveillance
“Now you will read a list of specific policy measures that some people have suggested the government should take on in response to threats against the United States. For each one, please select if you favor, oppose or neither favor nor oppose that particular policy.”
- Oppose warrantless listening to any telephone calls made in the U.S.66%
- Oppose warrantless monitoring of text messages and messaging apps63%
- Oppose warrantless reading of any emails sent in the U.S.62%
- Oppose warrantless monitoring of internet searches51%
- Favor warrantless listening to any telephone calls made in the U.S.14%
- Favor warrantless monitoring of internet searches27%
Subgroups (4)
- Republicans — Oppose warrantless monitoring of domestic phone calls74%
- Democrats — Oppose warrantless monitoring of domestic phone calls62%
- Republicans — Oppose warrantless monitoring of internet searches59%
- Democrats — Oppose warrantless monitoring of internet searches46%
U.S. adults · n=1,729 · ±3.2 · AmeriSpeak probability-based panel; online and telephone · 2021-08-12 – 2021-08-16 · source
Pew Research Center
Acceptability of police facial recognition in public spaces
“In your opinion, is it acceptable or unacceptable to use facial recognition technology in the following situations? [item: Law enforcement agencies assessing potential security threats in public spaces]”
- Acceptable — law enforcement assessing security threats in public spaces59%
- Not acceptable — law enforcement assessing security threats in public spaces15%
- Acceptable — landlords tracking who enters or leaves their buildings36%
- Acceptable — companies automatically tracking employee attendance30%
- Acceptable — advertisers seeing how people respond to public ad displays15%
Subgroups (4)
- Ages 18-29 — Acceptable (law enforcement, public spaces)42%
- Ages 65+ — Acceptable (law enforcement, public spaces)76%
- Black — Acceptable (law enforcement, public spaces)47%
- White — Acceptable (law enforcement, public spaces)64%
U.S. adults · n=4,272 · ±1.9 · American Trends Panel. Asked only of those who had heard of facial recognition (N=3,722); rebased to all U.S. adults · 2019-06-03 – 2019-06-17 · source
Pew Research Center
Trust in law enforcement to use facial recognition responsibly
“How much, if at all, do you trust the following groups to use facial recognition technology responsibly? [item: Law enforcement agencies]”
- Trust law enforcement agencies at least somewhat (NET)56%
- Trust law enforcement agencies a great deal17%
- Trust technology companies at least somewhat (NET)36%
- Trust advertisers at least somewhat (NET)18%
Subgroups (6)
- Ages 18-29 — Trust law enforcement at least somewhat49%
- Ages 65+ — Trust law enforcement at least somewhat67%
- White — Trust law enforcement at least somewhat61%
- Black — Trust law enforcement at least somewhat43%
- Rep/Lean Rep — Trust law enforcement at least somewhat65%
- Dem/Lean Dem — Trust law enforcement at least somewhat51%
U.S. adults · n=4,272 · ±1.9 · American Trends Panel, probability-based online panel. NOTE: asked only of the 3,722 who had heard of facial recognition; percentages are rebased to all U.S. adults, so they do not sum to 100 · 2019-06-03 – 2019-06-17 · source
Ipsos
Government use of AI and facial recognition to maintain order (26-country global total)
“Which of the following best reflects your opinion on the government's use of artificial intelligence and facial recognition to maintain order?”
- Should be allowed only under certain circumstances and subject to strict regulations65%
- Should be allowed as much as needed, even at the risk of citizens giving up privacy19%
- Should not be allowed under any circumstances16%
Online adults across 26 countries · n=19,106 · Ipsos Global Advisor. GLOBAL TOTAL ONLY — the U.S. row could not be read unambiguously from the published chart, so no U.S. figure is reported rather than risk misattribution · 2019-05-24 – 2019-06-07 · source
Gallup
Government collection of telephone and internet records
- Disapprove of the program53%
- Approve of the program37%
- Would be very concerned about their own privacy if government held logs of their calls or internet communications35%
- Edward Snowden did the right thing in sharing the information with the press44%
- Edward Snowden did the wrong thing42%
Subgroups (4)
- Democrats — Approve49%
- Independents — Approve34%
- Republicans — Approve32%
- Republicans — Disapprove63%
U.S. adults · n=1,008 · ±4 · Telephone interviews, landline and cell. Gallup does not print the verbatim topline in this release · 2013-06-10 – 2013-06-11 · source