Flock Was Searched More Than 100 Times After a Veteran Recorded Police
Flock became part of a federal civil rights case after Wisconsin deputies allegedly searched its camera network more than 100 times for one veteran’s vehicle. The searches followed Napoleon Jones recording a traffic stop and filing a complaint against the deputy who later arrested him. The number is striking, but the sequence matters more.
Jones was not publicly identified as a suspect in a new criminal investigation. According to his lawsuit, department personnel searched for his white BMW while reviewing his complaint against Deputy Brandon Shayhorn. Shayhorn, the subject of that complaint, conducted some of the searches himself.
The Waukesha County Sheriff’s Office says it cannot comment on pending litigation. Flock Safety also had not answered interview requests reported by local station TMJ4. The allegations therefore remain unresolved, and the defendants have not had their full case adjudicated.
However, records described in the complaint and deposition testimony raise a larger technology question. Flock markets automated license plate readers as tools for finding vehicles connected to crimes and missing people. What happens when the same searchable network becomes part of an internal effort to challenge the person complaining about an officer?
That conflict reaches beyond one Wisconsin parking lot. It tests whether access logs, written policies, and after-the-fact audits can restrain a surveillance system built for immediate, distributed searching.
A Recorded Stop Became a Searchable Trail
The dispute began with a phone recording, then expanded into a digital search for the recorder’s vehicle.
On May 4, 2025, Jones recorded Shayhorn conducting an unrelated traffic stop in Sussex, Wisconsin. Jones remained across the street and did not interfere, according to his federal complaint. The complaint identifies Jones as a Navy veteran and says the encounter later aggravated his diagnosed post-traumatic stress disorder.
After completing the original stop, Shayhorn entered the private parking lot where Jones had parked. The deputy later said Jones appeared suspicious and that his new BMW lacked a front license plate. Jones had a temporary registration displayed in the rear window.
Shayhorn activated his emergency lights while Jones was still inside the parking lot. Jones disputed the deputy’s authority to demand identification there and requested a supervisor. Shayhorn and another deputy, Brett Kromrey, removed Jones from the car, handcuffed him, and took him to jail.
Jones was held for about five hours, according to reporting based on the litigation. Authorities released him without a citation or criminal charge.
A supervisory memorandum dated May 6 plays a central role in the case. As quoted in the original complaint, a lieutenant told Shayhorn that the private parking lot did not permit enforcement of the suspected registration violation.
The memorandum listed limited exceptions, including impaired driving, reckless driving, fire-lane violations, and accessible-parking violations. Shayhorn reportedly confirmed that he had observed none of them. The supervisor concluded that Shayhorn lacked a lawful traffic stop and could not compel Jones to identify himself.
Jones filed a citizen complaint on May 5. That complaint challenged the stop and alleged retaliation for recording police.
The story might have remained a conventional dispute about a stop, arrest, and First Amendment activity. An image produced through an open-records request changed its scope.
The sheriff’s response contained a Flock image of Jones’ BMW dated May 30. The camera captured the vehicle roughly two miles from the parking lot and more than three weeks after the encounter, according to local reporting.
Attorney Nate Cade questioned why a complaint about a May 4 stop required a vehicle image captured 25 days later. A review of audit records then showed that several department members had searched for Jones’ vehicle more than 100 times.
The amended complaint does not treat those searches as a minor discovery dispute. It alleges that personnel used the network without a legitimate law enforcement purpose and in retaliation for protected recording activity.
That claim has not been decided. The records nevertheless show why searchable vehicle histories create a different category of institutional risk. A complaint that once involved reports, video, and witness accounts can now prompt repeated queries across a network of roadside cameras.
The searches did not merely revisit the location of the original encounter. They reportedly looked for Jones’ car on dates unrelated to the stop or the next day’s complaint.
That shift creates the article’s central tension. A system designed to help police investigate outsiders was allegedly turned toward someone asking the department to investigate itself.
The Searches Put Flock’s Guardrails on Trial
Flock can record each query, but a record of access does not establish that the access was justified.
An automated license plate reader, or ALPR, combines roadside cameras with software that converts vehicle images into searchable records. Flock’s system can store a plate number, vehicle image, make, color, time, date, and camera location.
An authorized user can search a known plate or broader vehicle characteristics. Agencies may also receive alerts when an observed vehicle matches a hot list or other qualifying record.
This capability changes the practical cost of surveillance. Following a car once required officers, time, and physical proximity. A database query can reconstruct sightings gathered by cameras that neither the searcher nor the target encountered directly.
Flock’s license plate policy says every query is logged with the username, time, purpose, plate, and other search elements. The company says those records support audits for legitimate access and compliance with law and agency policy.
The same policy places important responsibility on customers. Each agency designates administrators and authorized users. Customers decide how their data is shared, and agency policies determine how audits are conducted.
Flock encourages customers to adopt usage rules and regular auditing schedules. That language recognizes a basic limit: software can preserve evidence of a search, but agencies must decide who reviews it and what consequences follow.
Waukesha County’s policy prohibits personnel from using Flock data except for legitimate law enforcement purposes, according to TMJ4. The unresolved issue is whether investigating Jones’ temporary registration during a complaint review met that standard.
Deposition testimony reported by the station explains the department’s stated rationale. Shayhorn said Lieutenant Cory Farrell ordered him to search for the vehicle as part of determining whether Jones’ citizen complaint should be sustained.
Captain Lisa Panas also testified that her searches concerned the internal investigation. She said investigators wanted a photograph that might show whether Jones had displayed his temporary registration improperly.
That justification faces two immediate problems.
First, the reported searches reached dates other than May 4 and May 5. A camera image from another day could not establish how the temporary registration appeared at the precise moment Shayhorn activated his lights.
The amended complaint says Panas acknowledged that point during her deposition. A properly displayed tag on an earlier date would not prove proper display on May 4. An improperly displayed tag on another date would not establish the opposite.
Second, the deputy accused in the citizen complaint participated in the searches. When asked whether a deputy normally helps investigate a complaint against himself, Panas answered no. She also said she did not know why Shayhorn had been involved.
This is not simply an argument about how many searches are too many. One relevant, documented query can serve a legitimate investigation. A hundred poorly scoped searches can indicate that the investigative purpose was never adequately limited.
The audit trail made the alleged conduct discoverable. That is valuable. Without those records, Jones might have seen only the single May 30 image and never learned how often personnel had queried the system.
Yet discovery after litigation is a weak substitute for prevention. Jones and his lawyer had to notice an unexplained image, pursue records, and examine testimony before the search pattern became public.
Flock’s technical design therefore supplied both the contested capability and the evidence needed to challenge its use. That dual role matters, but it does not resolve who should have stopped the searches while they were occurring.
The Central Conflict Is Oversight Versus Access
A written restriction has limited value when the people conducting an investigation also define its legitimate purpose.
Police departments have credible reasons to want fast vehicle searches. ALPR systems can help locate stolen cars, identify vehicles connected to violent crimes, and find missing or endangered people.
Flock says its technology has helped solve approximately one million crimes and reunite roughly 10,000 missing people with their families. Those figures are company claims, not findings independently established within Jones’ litigation.
The Wisconsin case does not prove that every Flock search is unconstitutional or improper. It asks a narrower and more operational question: can agencies reliably prevent authorized users from turning a valid account into an invalid investigation?
The distinction between authorization and purpose is crucial. Shayhorn apparently had access to the system through his employment. The allegation is not that an unknown hacker breached Flock or stole credentials.
Instead, the claimed misuse occurred through normal institutional access. Deputies allegedly entered searches while performing work that supervisors connected to an internal complaint review.
That kind of use is harder to detect with conventional cybersecurity controls. Multifactor authentication can confirm who signed in. Role-based permissions can limit which employees reach the system. Encryption can protect records in transit and storage.
None of those controls can independently decide whether searching a complainant’s vehicle is necessary, retaliatory, or excessively broad. That judgment depends on policy, supervision, case documentation, and external review.
The reported rationale also reveals a conflict between investigative convenience and evidentiary relevance. A broad search can produce more images, yet more images do not necessarily answer the question under review.
If investigators wanted to assess the temporary plate on May 4, evidence from that place and time mattered most. According to the amended complaint, no Flock camera covered the parking lot where the arrest happened.
Searching other dates could reveal Jones’ later movements without resolving the original dispute. That is why the plaintiff characterizes the activity as an attempt to find a potential offense rather than investigate an existing one.
The defendants may challenge that characterization. They could argue that the searches were administrative fact-finding, that multiple query variations inflated the count, or that users were trying to locate a useful angle of the temporary tag.
The public record available through the reports does not fully explain how the total was counted. It also does not establish that every query returned a vehicle sighting. A search action and a successful observation are not identical.
Those uncertainties should limit the conclusion. More than 100 searches sound like continuous physical monitoring, but database queries can include retries, altered filters, different time windows, and unsuccessful requests.
Even with that caution, the personnel arrangement remains difficult to dismiss. Shayhorn testified that a lieutenant ordered his participation. Panas acknowledged that complaint subjects normally do not investigate complaints against themselves.
The institutional question is therefore larger than one deputy’s judgment. If supervisory direction produced the searches, the relevant control failure may sit within the review process itself.
A system cannot rely exclusively on agency managers to identify abuse if those managers authorize the disputed use. Independent auditing, defined case requirements, and timely anomaly detection become necessary precisely when internal incentives align.
That is the reversal at the heart of the case. The database was reportedly searched to evaluate a citizen’s allegation against government personnel, but the search itself became another allegation against those personnel.
New Controls Address the Pattern, Not the Verdict
Flock’s planned safeguards target abnormal access, but they cannot settle whether an officially ordered search is legitimate.
Flock announced platform changes in August 2026 amid growing public criticism and documented cases of unauthorized searches. The timing predates the broad national attention around Jones’ case, though it follows other reported misuse incidents.
According to an August platform update, law enforcement customers must implement an audit tool by January 1. Flock says the tool will identify abnormal search behavior and lock a user out pending internal review.
The company also plans to require users to connect searches to codes from their agencies’ records-management systems. Emergency overrides will be flagged for review.
Those changes speak directly to the mechanics exposed in Wisconsin. Repeated searches for one vehicle should become more visible. A required case code should also make it harder to conduct an undocumented query.
Flock is shortening its standard data-retention period from 30 days to seven. Evidence tied to a case number can remain available longer. Its June policy now describes a seven-day default, subject to customer law or policy.
Shorter retention reduces the historical window available for casual searches. It also limits the amount of location information exposed if an account or process is misused.
However, the Jones allegations show why mandatory case codes are not a complete answer. The searches reportedly related to an actual citizen complaint and an authorized internal investigation.
A user could potentially attach a valid complaint number to a search that remains disproportionate or irrelevant. The system would see documentation, while the central dispute would concern the meaning of that documentation.
An anomaly tool faces the same challenge. A hundred searches for one car might trigger review, but reviewers still need standards for evaluating necessity. If the searcher’s supervisors requested the activity, an internal review could reproduce the original decision.
Flock CEO Garrett Langley has said the changes should improve accountability. The company also says customers own their data and should set rules consistent with local community values.
Critics argue that local control becomes less reliable when agencies share access across jurisdictions. Data collected by one community can become searchable by users elsewhere, depending on permissions and applicable policies.
The scale has made Flock a political issue. The company says its cameras operate in 6,000 communities across every state except Alaska, according to national coverage.
That reach offers investigators a broader field of vehicle sightings. It also means that a local decision about one plate can reach far beyond the cameras purchased by the searcher’s own department.
Supporters see network effects that help police cross municipal boundaries. Critics see a national tracking infrastructure assembled through local contracts, often without a single national rule governing searches.
Wisconsin communities are already debating those tradeoffs. In March, Dane County supervisors introduced a resolution seeking to remove funding for the sheriff’s Flock system.
The county said the sheriff’s office had granted access to more than 140 agencies across at least eight states. Public audit data showed that a majority of recent deputy searches used the broad “attempt to locate” category.
The proposed funding resolution cited concerns about data sharing, constitutional rights, auditing, and uses beyond the system’s original purpose. It also acknowledged continuing disagreement about the tool’s public-safety benefits.
These debates are not exact parallels to Jones’ case. Dane County involves a different agency, contract, and policy process. The comparison matters because it shows that the pressure on Flock is no longer limited to hypothetical privacy concerns.
Communities now have concrete categories of conduct to evaluate: searches involving romantic partners, immigration enforcement, complaints against officers, and broad investigative labels. Each category tests whether product controls can distinguish public safety from institutional convenience.
Flock’s new restrictions should reduce certain forms of casual misuse. They do not resolve the constitutional questions surrounding aggregated location histories. They also do not determine whether a search connected to an internal investigation becomes retaliation.
Those answers will come from courts, legislation, contract terms, and agency discipline. Product changes can support those institutions, but they cannot replace them.
Three Signals Will Show Whether Accountability Catches Up
The next phase depends on the federal case, the implementation of mandatory controls, and decisions by local governments.
The first signal is the progression of Jones’ lawsuit in the Eastern District of Wisconsin. His amended complaint, filed July 31, 2026, added personnel allegedly involved in the Flock searches and expanded the retaliation claims.
The filing alleges violations involving the stop, arrest, search activity, inadequate training, and a conspiracy to deprive Jones of constitutional rights. These remain allegations, and the court has not issued a final decision on their merits.
Future motions could clarify whether the court treats the ALPR activity as a Fourth Amendment search. They could also address whether repeated queries support a First Amendment retaliation claim when officials connect them to complaint review.
A ruling favorable to Jones would strengthen the argument that agencies need stricter limits around complainants and other people not suspected of crimes. A dismissal or narrower ruling would weaken broader constitutional claims without necessarily validating the department’s practices.
The second signal is how Flock’s mandatory controls work after January 1. The useful metric will not be whether agencies enable an audit feature. It will be whether the feature detects questionable patterns early and produces independent action.
Important questions include how the system defines abnormal behavior, who receives an alert, and whether a supervisor can clear it without outside review. Agencies should also disclose how many searches trigger lockouts and how often reviews find policy violations.
Case-number requirements deserve similar scrutiny. A case identifier should represent more than an administrative field. Agencies need rules connecting each query to a specific investigative necessity, offense category, time range, and authorized purpose.
The third signal is the response from city councils, county boards, and state lawmakers. Local governments control many contracts and can require narrower sharing, shorter retention, warrant rules, public reports, or independent audits.
Some communities will keep the cameras while adding restrictions. Others will suspend or end contracts. Those decisions will reveal whether Flock’s product changes have restored confidence or merely shifted responsibility back to customers.
The Jones case should also change what journalists and residents request through public-records laws. A single vehicle image may tell only a fraction of the story. Query logs, case codes, time windows, user identities, supervisory approvals, and audit findings provide the fuller picture.
For technology buyers, the practical lesson is equally direct. Logging is not governance. A platform can record every action and still permit a harmful pattern until someone with authority recognizes it.
Procurement reviews should test internal misuse scenarios, not only external breaches. Buyers should ask what happens when an authorized employee searches a neighbor, former partner, critic, journalist, complainant, or colleague.
They should also examine whether the person reviewing an alert is independent from the underlying dispute. Escalation rules matter when management, rather than a rogue employee, approves the contested action.
Flock’s network illustrates a broader problem with searchable institutional data. Collection happens continuously, while scrutiny often begins only after a disputed search becomes visible.
That imbalance can affect location systems, workplace monitoring, communications archives, and other databases. The risk grows when ordinary employees can retrieve historical records without establishing relevance before access.
Jones discovered the searches because one unexplained image appeared in a records response. Many people would never know that their vehicle had become an internal search subject.
His case has not established that the searches were unconstitutional retaliation. It has established a detailed public dispute about who searched, who ordered the searches, what rationale they offered, and why the subject deputy participated.
The final question is not whether police should possess useful investigative technology. It is whether usefulness earns access before necessity, proportionality, and independent oversight have been demonstrated.
Readers should watch the court docket, Flock’s post-January audit results, and local contract votes. Those three signals will show whether the response produces enforceable limits or another layer of policy language.
The more than 100 Flock searches matter because they turn an abstract surveillance debate into an auditable sequence of decisions. Follow those decisions, ask who approved them, and demand evidence that the next questionable search will be stopped before litigation exposes it.
