What the Evidence and the Record Show About Flock's Cameras

LANSING, Mich. — The Lansing Committee on City Operations takes up the city's Flock Safety deployment at 4:00 PM today, Thursday, April 23, 2026. Two other things the committee will need in front of it: what the independent research says about whether automatic license plate reader (ALPR) networks prevent crime, and what the 2024 to 2026 public record shows about Flock's data in other jurisdictions. The ownership of the vendor and the structure of the "Flock Grant system" named on the agenda are covered separately in Who Owns and Funds the Cameras Watching Lansing.

Both records are public. The research record sits in peer-reviewed journals and federal policy reports; the deployment record sits in court dockets, state-agency audits, and civil-liberties reporting. The comparison that the committee's review can make is between the safeguards described for Lansing's deployment and the harms documented in other Flock jurisdictions.

What the peer-reviewed research says about crime reduction

The Center for Evidence-Based Crime Policy (CEBCP) at George Mason University hosts a public research portal on police technology evaluation that links to the Lum and Koper ALPR field trials and related studies. Those studies characterize the evidence base as mixed, with impact dependent on deployment scale, the information supporting the LPR system, and the specific ways LPRs are used.

A 2025 study by Shjarback and Sarkos, published in Justice Evaluation Journal, evaluated a major ALPR expansion in one city. The study reports a result mixed across crime categories: the expansion did not reduce violent crime overall, but shootings (a subcategory of violent crime), motor-vehicle thefts, and property crime all fell after the deployment. The authors are careful about cause and effect. Shootings and the other crime categories went down in the same period that the cameras were being rolled out. That does not by itself prove the cameras caused the decline. Many other things change in a city from one year to the next (policing tactics, drug markets, weather, demographics), and a study of one city cannot rule out those other explanations. What the study can say is what did and did not go down after the cameras went in.

Earlier studies using the strongest available research designs (randomized trials, and comparisons of neighborhoods with cameras to similar neighborhoods without them) in Mesa, Arizona, Alexandria, Virginia, Fairfax County, Virginia, and Baton Rouge, Louisiana found that ALPRs did not significantly reduce crime in the areas where they were deployed. One study, in Buffalo, New York, did find a reduction in the FBI's serious-crime categories (murder, rape, robbery, aggravated assault, burglary, larceny, motor vehicle theft, and arson) when ALPRs were paired with vehicle checkpoints. The summary and citations are collected in the Shjarback and Sarkos evaluation linked above.

The National Institute of Justice's summary of the Koper and Lum 2019 study notes that the share of auto-theft and robbery cases that police closed (what researchers call the clearance rate) rose after the cameras were deployed, but the rise was not large enough to rule out chance once the study accounted for other factors that change year to year. The authors cautioned that "other factors may have also contributed." Agency self-reports diverge from the peer-reviewed evaluations. A 2012 survey funded by the National Institute of Justice and run by Roberts and Casanova at the International Association of Chiefs of Police (responses from 305 agencies, with follow-up data from 40 ALPR-using agencies) found that 68 percent of responding ALPR users self-reported increased stolen-vehicle recoveries, 55 percent reported increased arrests, and 50 percent reported increased productivity, per the abstract in the federal National Criminal Justice Reference Service. A separate NIJ synthesis notes that "rigorous evaluation of LPRs has been very limited."

Work by researchers from the RAND Corporation (a nonprofit policy research institute) published by NIJ concluded that ALPR "appears to provide utility to any type of investigation if supported by appropriate data access and retention policies," framing the benefit as dependent on policy controls rather than on the technology itself. The International Association of Chiefs of Police's March 2018 Research in Brief summarizes the consistent finding across the peer-reviewed literature: ALPRs increase arrests and stolen-vehicle recoveries during patrols where they are deployed, by a factor of roughly two to three times, but have not produced clear evidence that they deter crime from happening in the first place.

The one independent test of Flock's cameras

IPVM, a subscription-funded surveillance-technology testing publication, reported in 2021 that its test of a Flock Falcon camera measured approximately a 10 percent error rate, with misclassification particularly common for license-plate state, vehicle type, and vehicle make (including one documented instance where the camera classified a yellow Nissan Xterra as a bus). The full IPVM test methodology is behind a subscription paywall; the findings have been widely cited in secondary coverage but cannot be independently verified by a reader without an IPVM subscription. Flock halted sales to IPVM after publication and has since refused to make cameras available for independent re-testing.

Flock markets its technology as "best-in-class, consistently performing above other vendors." The American Civil Liberties Union has urged communities to "reject surveillance products whose makers won't allow them to be independently evaluated," citing Flock's refusal to make cameras available to independent testers. No dedicated accuracy benchmark for municipal ALPR deployments has been published by an industry standards body (for example, the National Institute of Standards and Technology or the Institute of Electrical and Electronics Engineers) that could be located in this reporting. ALPR accuracy in the literature is typically reported per-vendor and per-field-condition, not against an industry-standard benchmark.

What the record shows Flock has done with the data

Cross-customer "National Lookup." Flock's National Lookup feature was reportedly disabled in 2023 to comply with California SB 34 (the state's ALPR Privacy Act, which limits how police can share plate-reader data with other agencies), and was later found to be active on specific deployments without customer consent. Santa Cruz, California Police Chief Bernie Escalante identified violations of California SB 34 and SB 54 via national searches that continued until February 11, 2025, per Lookout Santa Cruz. Mountain View, California's first Flock camera had National Lookup enabled from August to November 2024 without city authorization, per Peninsula Press. Santa Cruz Council voted six to one to end its contract, and Mountain View voted to terminate in February 2026.

Illinois Secretary of State audit. An August 2025 audit of twelve Illinois agencies by Illinois Secretary of State Alexi Giannoulias found Flock had facilitated U.S. Customs and Border Protection access to Illinois ALPR data in violation of the 2023 Illinois statute barring ALPR use for immigration or abortion enforcement, per the Secretary of State's August 25, 2025 press release. Flock paused its federal-agency pilot program nationwide in response, and the Secretary of State's office stated that the investigation remains open.

Federal data-sharing without customer consent. An October 2025 report from the Center for Human Rights at the University of Washington documents that CBP accessed ALPR data from Flock customers who had not intended to share with federal authorities, via an undisclosed pilot program. In Washington State, at least eight local agencies enabled direct sharing with Border Patrol, and at least ten more had "back door" access without having authorized it.

California class action. A state-law class action filed in San Francisco Superior Court in February 2026, per the plaintiffs' counsel announcement from Gibbs Mura and Milberg, alleges Flock shared California plate data with federal agencies (Immigration and Customs Enforcement (ICE); Customs and Border Protection (CBP); the FBI; and the Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF)) 1.6 million times in violation of California's ALPR Privacy Act. The filing venue and date are drawn from the plaintiffs' law-firm press release; no court docket has been independently pulled for this post.

Condor camera live-feed exposure. In January 2026, security researchers found at least sixty Flock Condor cameras streaming live with unencrypted, unauthenticated admin interfaces; archival footage and deletion controls were externally accessible, per WFLX reporting. Flock publicly disputed the characterization as a "breach."

Stolen customer credentials. TechCrunch reported in November 2025 that Flock customer logins had been found on underground sites where credentials stolen by malware from infected computers are traded (what security researchers call stealer-log marketplaces). Multi-factor authentication (the second login check that asks for a code from a phone or token in addition to a password) had not been mandatory on Flock accounts at the time of the reporting. Flock enabled multi-factor authentication by default for new customers starting November 2024, and told TechCrunch that by November 2025 approximately 97 percent of existing customers had turned it on.

Litigation, regulatory actions, and pending cases

The core legal question is whether mass, continuous ALPR collection counts as a Fourth Amendment search (a government search that would normally require a warrant). So far, the rulings on that question go two to nothing for cities and their Flock deployments. On January 27, 2026, Judge Mark Davis of the U.S. District Court for the Eastern District of Virginia ruled for the city in Schmidt v. City of Norfolk on summary judgment (a ruling that decides the case without a trial because no material facts are in dispute), holding that Norfolk's 176-camera Flock deployment does not count as a Fourth Amendment search at current density. A 2024 Norfolk Circuit Court ruling that had gone the other way was reversed by the Virginia Court of Appeals. The Institute for Justice has appealed Judge Davis's ruling to the Fourth Circuit. No appellate court has yet ruled against an ALPR deployment on the merits. Judge Davis noted in the same opinion that "as the number and capabilities of ALPR cameras expand, the constitutional balancing could conceivably tip the other way," per Courthouse News.

Parallel cases are pending in other courts:

• San Jose federal class action, Northern District of California, April 2026. IJ, on behalf of three San Jose residents, challenges 474 Flock ALPRs citywide and the more than 1,000 San Jose Police Department employees who can search the data without a warrant and without any particular reason to suspect the driver, per the EFF press release.
• Moore v. City and County of San Francisco, 3:25-cv-11011, Northern District of California, filed December 28, 2025. Fourth Amendment and California SB 34 claims. The complaint alleges San Francisco Police Department (SFPD) data was accessed by out-of-state agencies from August 2024 through February 2025, with at least nineteen searches flagged "related to Immigration and Customs Enforcement," per the Civil Rights Litigation Clearinghouse.
• People of California v. City of El Cajon, October 2025. California Attorney General Rob Bonta filed enforcement action against El Cajon over Flock data sharing with more than 100 out-of-state agencies in violation of SB 34, with a hearing scheduled, per the California Attorney General's press release.
• May v. City of Eugene, Lane County Circuit Court, Oregon, filed October 2025. An ACLU of Oregon Public Records Act challenge after the city refused to disclose approximately sixty Flock camera locations, per the ACLU of Oregon case page.
• Washington State Public Records ruling, November 6, 2025. Skagit County Superior Court held Flock images are public records under the Washington Public Records Act, per the EFF summary. Five cities and one county deactivated cameras in response. Washington SB 6002, signed by Governor Bob Ferguson in March 2026, largely exempted ALPR images from the PRA, and a motion to vacate the November ruling was denied.
• Congressional investigation and FTC referral. An August 6, 2025 letter from Representatives Raja Krishnamoorthi and Robert Garcia to Flock CEO Garrett Langley demanded records of national-lookup searches tied to ICE, CBP, and abortion enforcement. A November 3, 2025 joint letter from Senator Ron Wyden and Representative Krishnamoorthi to Federal Trade Commission Chair Andrew Ferguson requested an investigation of Flock's cybersecurity and authentication practices; no public confirmation that the FTC has opened a formal investigation has been located for this post.

Wrongful detentions tied to Flock hits

Brandon Upchurch, Toledo, Ohio, April 2024. A Flock camera misread a "7" on a plate as a "2" and produced a stolen-vehicle alert; a K-9 was deployed at the resulting stop, and Upchurch was bitten, hospitalized, and jailed. In October 2025 a federal court approved a $35,000 settlement, and the judge reportedly remarked during the settlement conference that "Flock Flocked up," per the WTOL report.

Chrisanna Elser, Columbine Valley, Colorado, October 2025. A Flock hit was used by an officer to accuse Elser of a $25 package theft, and the officer refused to show Elser the footage of the alleged theft. Elser produced independent dashcam footage showing Elser was elsewhere at the time of the alleged theft, per Colorado Sun reporting. No charges were filed, and no civil suit was filed at the time of publication of this post.

Redmond, Washington, father. The camera read the plate correctly. The plate was registered to the father; the felony warrant was attached to the son, who has a similar name. The system matched the plate to the warrant without distinguishing the registered owner from the person actually wanted. The father was handcuffed in the family driveway. The case, reported by KING 5, illustrates that a correctly-read plate can still produce a wrongful detention when the list of "wanted" records the camera is matched against does not distinguish between the owner of a vehicle and a different person with a similar name.

Sherwood, Arkansas, family of four, March 2026. An ALPR misread caused by a plate-frame obstruction; the family was ordered from the vehicle at gunpoint. Source: Carscoops.

Officer misuse for personal reasons

Every one of the cases below was surfaced because Flock's audit logs captured the queries in a form investigators could later reconstruct. Flock Safety spokesperson Holly Beilin told The Marshall Project that the logs "represent a key accountability feature, because they can't be changed after the fact." The logs record what happened, but they do not stop it from happening: Flock's system does not require a supervisor to approve a search before it runs, and does not tell the owner of a plate that the plate has been searched.

• Lee Nygaard, former Chief of Police, Sedgwick, Kansas, 2024. Nygaard ran 228 total Flock queries tied to personal relationships (164 searches of an ex-partner's plate and 64 of that ex-partner's new partner's plate) and also followed the couple out of town in a department cruiser, per KAKE reporting. Disposition was eighteen months probation, loss of police certification, and resignation; criminal charges were not pursued.
• Josue Ayala, Milwaukee Police Department, charged February 2026. Ayala ran 124 Flock searches of a partner's plate and was charged with one count of misdemeanor misconduct in public office, per Wisconsin Examiner reporting.
• Former Costa Mesa, California officer. Pleaded guilty after using Flock over nine months to track a spouse, another romantic partner, and rivals, per CBS LA reporting.
• Sandy Springs, Georgia officer. Resigned after internal affairs sustained findings of "Abuse of Position and Authority" and "Conduct Unbecoming" for using a department Flock login to pull data for a private contractor, per Appen Media reporting.
• Menasha, Wisconsin officer. Pleaded not guilty in 2026 to a charge of using Flock to track an ex-partner, per The Marshall Project reporting.

Protest and abortion-travel surveillance

The Electronic Frontier Foundation obtained datasets of more than twelve million Flock searches from approximately 3,900 agencies covering December 2024 through October 2025. The EFF analysis documents that more than fifty federal, state, and local agencies ran searches tied to protest activity; nineteen agencies conducted searches specifically tied to "No Kings" protests in June and October 2025.

The Michigan entry in that dataset is Grand Rapids Police Department. GRPD ran five Flock searches tied to a "Stand Up and Fight Back Rally" in February 2025, with the reason field logged simply as "Protest," hitting approximately 650 networks, per Techdirt's summary of the EFF findings.

In May 2025, a deputy in Johnson County, Texas searched more than 83,000 Flock cameras across 6,809 networks, including cameras in Washington and Illinois (states where abortion is legal), to locate a woman who had self-managed an abortion. The deputy's initial search reason read: "had an abortion, search for female." The sheriff publicly framed the search as a welfare check. Court records later showed deputies had opened a "death investigation" and consulted prosecutors about charging the woman, per the May 2025 EFF reporting and the October 2025 EFF follow-up.

The evidence and the record, laid out

Claim	Public Record
ALPR reduces crime	Multiple randomized and quasi-experimental evaluations found no statistically significant reduction in overall crime. The 2025 Shjarback and Sarkos evaluation in Justice Evaluation Journal reported mixed results: no reduction in violent crime overall; reductions in shootings, motor-vehicle theft, and property crime. See also IACP Research in Brief, March 2018.
Flock's cameras are highly accurate	The only independent test of a Flock Falcon camera (IPVM 2021, subscription) measured a 10 percent error rate in camera readings. Flock halted sales to IPVM after publication and has refused independent re-testing since. No industry-standard accuracy benchmark exists.
Data does not go to the feds without local permission	Illinois Secretary of State August 2025 audit finding; October 2025 University of Washington report documenting CBP back-door access in Washington State; February 2026 California class action alleging 1.6 million out-of-state queries.
Officers cannot misuse the system	One former police chief and at least four additional officers across five states charged or disciplined in 2024 to 2026 for using Flock to stalk romantic partners or pull data for private contractors (KAKE; Wisconsin Examiner; CBS LA; Appen Media; The Marshall Project). Audit logs surfaced every cited case after the fact; no cited case was caught before the misuse was complete.
The cameras are not used to surveil protesters	More than fifty agencies ran protest-related searches in the EFF-obtained dataset, including Grand Rapids PD in Michigan (Techdirt summary).
Courts have upheld the technology	The merits record on Fourth Amendment theories to date is 2-0 for ALPR defendants: federal Norfolk summary judgment (Courthouse News) and the Virginia Court of Appeals reversal of a trial-court suppression (GovTech). Multiple other cases are pending. The Norfolk federal ruling is under appeal.

The Lansing record

The Flock discussion is Item 5.C on the Committee on City Operations agenda for April 23, 2026, listed as "DISCUSSION: Update from LPD on Flock Grant system and safeguards in place." It is one of several items on the agenda. The committee packet contains no written materials accompanying the item. The scope of the discussion, the content of any verbal update, and the depth of any question-and-answer with members are not determinable from the agenda itself.

No court or regulator has made a finding that the Lansing Police Department has operated the Flock deployment unlawfully. No independent audit of the Lansing deployment has been published. No written Lansing-specific Flock policy or data-retention rule (the document setting how long plate data is kept) has been located in any published city document.

The records that would let residents compare the Lansing deployment against the deployments in other cities described above are the Flock contract, any written policy, the data-retention rule, the audit-log summary, any grant documents, the platform's configuration settings, and the records showing who is allowed to run searches. None of those records has been posted to the city website or appeared in a CivicClerk committee packet at time of publication.

What this post claims and what it does not

This post does not claim that ALPR technology is useless to investigations; the academic record is explicit that ALPR increases arrests and stolen-vehicle recoveries during patrols where the technology is deployed. It does not claim that Flock has operated unlawfully in Lansing, because no court or regulator has made such a finding about the Lansing deployment. Nor does it claim that every officer who uses a Flock system will misuse it; the cited misuse cases involve individual users, and Flock's audit logs surfaced each case after it happened. And it does not claim that the Norfolk ruling is a settled national precedent, because that ruling is under appeal.

What this post does document is the peer-reviewed evidence on crime reduction as of publication, the accuracy test record, and the 2024 to 2026 record of Flock deployments in other jurisdictions. Whether any of that record surfaces at the April 23 committee meeting is a matter for the meeting itself.

Also in this series

• Lansing ALPR Map. Interactive plot of 88 automatic license plate reader cameras and 104 CCTV cameras across the greater-Lansing area, filterable by operator and by Flock manufacturer. Built from OpenStreetMap; each pin links back to the underlying OSM record.
• Why Does Lansing Need Flock?. A Lansing City Council committee reviewed the LPD Flock deployment in public for the first time on April 23, 2026. The accountability chain up to that meeting, and six questions the committee's review should produce answers to.
• Who Owns and Funds the Cameras Watching Lansing. The vendor behind the cameras, the private-capital backing, the state master contract Michigan agencies buy under, and what the "Flock Grant system" named on the committee agenda actually describes.

---

Sources

Academic and policy research is drawn from the Center for Evidence-Based Crime Policy LPR portal (George Mason University), the 2025 Shjarback and Sarkos ALPR expansion evaluation in Justice Evaluation Journal, National Institute of Justice summaries of Koper and Lum 2019 (NIJ), the NIJ effectiveness synthesis (NIJ), the 2012 Roberts and Casanova IACP / NIJ-funded ALPR implementation survey (NCJRS abstract), RAND-affiliated work published as NIJ report 247283, and IACP Research in Brief, March 2018. The independent accuracy test of Flock cameras is IPVM 2021; ACLU commentary on Flock's refusal of independent testing is at the ACLU 2024 statement. The Illinois Secretary of State audit finding is Giannoulias, August 25, 2025; the Forest Park follow-up is Giannoulias, September 5, 2025. Federal-agency back-door access is documented in the University of Washington Center for Human Rights report, October 2025. The California class action is Gibbs Mura and Milberg announcement. The National Lookup terminations are Lookout Santa Cruz and Peninsula Press. The Condor live-feed reporting is WFLX, with Flock's response at Flock's blog. Credential reporting and MFA-adoption details are TechCrunch. The Norfolk federal summary judgment is reported at the IJ case page, Courthouse News, and WHRO. The San Jose IJ class action is documented at EFF and the IJ case page. Moore v. CCSF docket and complaint are at the Civil Rights Litigation Clearinghouse. People of California v. City of El Cajon is at the California AG press release. May v. City of Eugene is at the ACLU of Oregon case page. The Washington Public Records ruling is at EFF. Congressional and FTC letters are at Rep. Krishnamoorthi's press release and the Wyden letter PDF. The Upchurch settlement is at WTOL. The Elser case is at the Colorado Sun. The Redmond WA case is at KING 5. The Sherwood AR case is at Carscoops. Officer-misuse cases are at KAKE (Nygaard), Wisconsin Examiner (Ayala), CBS LA (Costa Mesa), Appen Media (Sandy Springs), and The Marshall Project (Menasha and others). Protest surveillance is at EFF, November 2025, with the Grand Rapids detail at Techdirt. Johnson County TX abortion surveillance is at EFF, May 2025 and the EFF October 2025 follow-up. Items flagged as unverified or not attributable to Flock specifically (the widely-cited Brittney Gilliam Aurora case; "Contee Chatman" Chicago case; "Overland Park" or "Greenville NC" cases; the "35 stolen credentials" specific count; the June 2025 "toggle flip" date; Michael Steffman arrest specifics) are addressed in the underlying research memos and are not cited in this post.