By Brandon Burley

KNOXVILLE, Tenn. — A series of reports to the nation’s clearinghouse for suspected online child exploitation helped Knoxville police identify a suspect, search a Gleason Drive home and make an arrest last week.

But when The Redemption Project asked what happens to the thousands of other CyberTips entering Tennessee’s Internet Crimes Against Children system, Knoxville police disclosed a potentially more consequential number.

There are 190 actionable tips waiting to be assigned.

The oldest dates to October 2025.

“Our current backlog of actionable tips awaiting assignment is 190 with the oldest tip being October 2025,” Knoxville Police Department Communications Manager Scott Erland told TRP in a written response Sept. 1.

There is an important limitation to that disclosure.

KPD has not yet clarified whether those 190 actionable tips are waiting inside the Knoxville ICAC unit, sitting in a statewide Tennessee ICAC queue before assignment to an affiliate agency, or represent another defined workload category.

TRP asked KPD to make that distinction Tuesday. No additional response had been received as of Wednesday morning.

The number also does not mean 190 children are known to be in immediate danger while police wait to respond.

KPD defines an “actionable tip” as one containing information related to an actual crime and enough “solvability factors” for investigators potentially to identify and locate a suspect or victim. Erland said threats involving imminent harm trigger immediate notification and response.

That distinction matters.

A CyberTip is not automatically a crime. It is not necessarily a unique offender. It is not an arrest, and it is not necessarily a child waiting for rescue.

But a lead that police themselves have classified as actionable and that is still awaiting assignment is a different measurement from the millions of raw reports entering the national system.

And some of Tennessee’s oldest actionable work is now old enough that KPD’s description reaches back to last October.

One 2025 number does not reconcile

KPD supplied another figure that TRP is not prepared to publish without qualification.

Erland said the Tennessee ICAC Task Force received 37,755 CyberTips in 2025, with 2,298 unique investigations during the year. He said duplicate CyberTips are not formally tracked but are “exceedingly rare.”

The first number presents an unusual problem.

The National Center for Missing & Exploited Children’s official 2025 CyberTipline Reports by State table says 49,541 reports resolved to Tennessee: 35,504 classified as referrals and 14,037 classified as informational reports.

The same NCMEC table lists Indiana at exactly 37,755 reports — 28,000 referrals and 9,755 informational reports.

That does not prove KPD gave TRP the wrong state’s number.

The two organizations may be measuring different stages of the system. NCMEC’s table counts reports that resolved to a state, while KPD may be describing a narrower universe that actually entered or moved through the Tennessee ICAC workflow.

But because the number KPD supplied matches Indiana’s official total exactly, TRP asked the department whether 37,755 was inadvertently transposed from the wrong state or whether it describes a specific operational subset of Tennessee reports.

Until KPD answers that question, 37,755 should not be treated as Tennessee’s definitive 2025 CyberTip total.

Even the word “referral” means different things

There is another reason the numbers cannot simply be placed side by side.

KPD and NCMEC use some of the same words differently.

In its written response, KPD told TRP that a referral is a complaint received from or forwarded by another law-enforcement or government agency rather than NCMEC. A CyberTip, KPD said, is an investigative lead provided through NCMEC.

NCMEC uses “referral” differently in its national reporting.

It calls a CyberTipline report a referral when the submitting company has provided sufficient information for law enforcement potentially to act, usually including information about a user, imagery and a possible location for the child or offender.

NCMEC calls a report informational when the information is insufficient for that purpose or when, among other things, imagery has been widely circulated and repeatedly reported.

Those classifications describe the quality and usability of information reaching NCMEC.

They are not necessarily equivalent to KPD’s definition of an actionable tip, a referral or a unique investigation.

That makes the 2,298 unique investigations KPD reported potentially more useful than a raw CyberTip total — but even that number needs to be understood within KPD’s own case-definition system.

TRP therefore is not calculating a percentage of reports that became investigations. The numerator and denominator have not yet been shown to measure compatible things.

The national volume is enormous — and raw totals can mislead

NCMEC received approximately 21.3 million CyberTipline reports in 2025.

It identified and escalated more than 53,000 reports as urgent or involving a child in imminent danger. More than 2 million reports resolved to the United States, including about 1.9 million that resolved to a specific state.

Those figures sound overwhelming because they are.

They also require context.

Electronic service providers submit the vast majority of CyberTipline reports, and reported files can include imagery that has circulated repeatedly for years. NCMEC uses hash matching and other triage tools in part to reduce duplication at the file level and help investigators prioritize newer or more urgent material.

A report can also be made available to more than one law-enforcement agency if it contains geographic connections to different jurisdictions. NCMEC expressly warns that a single report can have multiple geographic nexuses.

That is why a headline based solely on the largest available CyberTip number can tell readers surprisingly little about the actual investigative workload.

The 190 figure is different.

KPD says those tips have already been classified as actionable and are waiting for assignment.

Tennessee was warned about this problem in 2024

The current disclosure is not the first indication that Tennessee’s child-exploitation investigators have struggled with volume.

An October 2024 Tennessee Comptroller performance audit of the Tennessee Bureau of Investigation devoted an entire observation to the issue.

Auditors warned that CyberTips related to internet crimes against children would likely continue to rise, “creating a backlog and overburdening the bureau’s limited investigative resources.”

At TBI, the word “backlog” had a specific meaning.

The agency classified a CyberTip as being on “active hold” until an agent became available to review it. As of April 1, 2024, the TBI ICAC squad had 2,543 cases in that status.

Auditors found that the bureau’s ICAC squad still had the same four agent positions it had when the squad was created in 2007.

From June 2020 through March 2024, TBI opened 3,044 cases related to CyberTips involving child sexual abuse material. Auditors found 84% contributed to the backlog and 90% involved allegations classified as child exploitation.

“The Internet Crimes Against Children Squad’s staffing levels have not kept pace with the growing backlog of cybertips,” auditors wrote.

TBI agreed with the observation and said it would continue seeking additional positions.

That was a TBI backlog, not the current 190-tip queue KPD has disclosed.

The numbers should not be added together or treated as the same measurement.

What they establish is continuity in the systems question: Tennessee auditors documented an investigative-assignment problem in 2024, and KPD is now reporting actionable work still awaiting assignment in 2026.

KPD sits at the front of Tennessee’s routing system

Tennessee’s ICAC structure makes the scope of the 190 especially important.

The 2024 Comptroller audit explains that NCMEC electronically transmits a Tennessee CyberTip to the state ICAC commander based on where NCMEC believes the victim or potential offender is located.

In Tennessee, that commander works at KPD.

KPD then triages the report and routes it to an affiliate law-enforcement agency covering the relevant geography. When no local affiliate covers a jurisdiction, KPD can assign the tip to TBI.

That creates at least two conceptually different places where something could be “awaiting assignment.”

A report could be waiting for routing to the law-enforcement agency that will own it.

Or an affiliate could already have received it and be waiting to assign the matter to an individual investigator.

KPD’s initial response does not distinguish between those possibilities for the 190.

That is why the answer to TRP’s follow-up question could materially change the headline’s statewide significance.

The task-force network has grown

The statewide network itself is larger than KPD’s public webpage suggests.

The department’s current ICAC page still says KPD works with 60 statewide affiliate agencies and lists the KPD unit as one lieutenant, one sergeant, three investigators and two computer forensic examiners.

KPD’s Aug. 28 Gleason Drive release said the task force links 80 state, local and federal agencies.

Erland told TRP the current number is actually 82 affiliated agencies when KPD is included, explaining that the task force has grown since the webpage was updated.

He also supplied an updated KPD staffing configuration: three detectives, three digital forensic analysts and two supervisors assigned to the ICAC unit.

That is eight people.

KPD said the Tennessee task force has approximately 164 members across participating agencies.

Those 164 people should not automatically be treated as 164 full-time ICAC investigators. Participating agencies can assign personnel in different capacities and at different levels of time commitment.

The 2024 audit illustrates that distinction. It calculated designated ICAC staffing in full- and partial-time equivalents rather than simply counting everyone connected with an affiliate.

One queue KPD says is not backed up

KPD’s answers also identify an important place where the department says there is not a backlog.

Asked how many devices were awaiting forensic examination, Erland told TRP:

“We have no forensic exam backlog at KPD. Turnaround time is case-dependent.”

That matters because digital child-exploitation investigations can involve two separate capacity problems.

Investigators first have to receive, triage and work the investigative lead.

After devices are seized, forensic personnel may have to extract and analyze large amounts of digital evidence.

KPD’s current disclosure puts the reported backlog at the actionable-tip assignment stage, not in its local forensic-examination queue.

That is useful operational context — and another reason the 190 should be defined precisely before being compared with other state or federal backlog figures.

The Gleason Drive case shows what happens when a CyberTip becomes a case

The reporting began with an Aug. 28 operation at a residence on Gleason Drive.

KPD said its ICAC detectives received a series of NCMEC CyberTips containing information concerning alleged exchanges of child sexual abuse material and alleged inappropriate communications with children through multiple online platforms.

Investigators identified Yaac Pekle Mendez, 33, searched the residence, recovered evidence and arrested him, according to KPD.

Mendez was booked on three counts of sexual exploitation of a minor.

The investigation remains active. The charges are allegations, and Mendez is presumed innocent unless and until proven guilty.

KPD declined to answer TRP’s case-specific questions about when the underlying reports arrived, whether any related to the same underlying conduct, what investigative steps occurred between receipt of the reports and the residential search warrant, and whether preservation demands, subpoenas, provider warrants or other legal process preceded the search.

Erland cited the continuing investigation.

The department did answer the broader systems questions.

KPD defined a CyberTip as an investigative lead supplied through NCMEC; an investigation as an agency’s investigative response to a complaint or lead; a child identified as a victim whose identity was previously unknown to law enforcement but discovered during an investigation; and a child rescue as locating a child and removing that child from ongoing physical or sexual abuse.

Those definitions help prevent raw activity statistics from becoming something they are not.

A million reports are not a million offenders.

A CyberTip is not an arrest.

A backlog does not establish that imminent-danger reports are sitting unattended.

But the inverse matters, too.

An actionable lead that law enforcement says contains evidence of an actual crime and enough information potentially to identify or locate a suspect or victim — yet remains unassigned — is a public-safety workload measure worth understanding.

KPD says there are 190 of them.

The oldest dates to October 2025.

The remaining question is where, exactly, those 190 cases are sitting — and how long actionable leads ordinarily wait before an investigator takes ownership of them.

TRP has asked KPD to answer that question and will update this report when it does.

I am a retired detective and criminal justice / government educator based in Tennessee. I founded The Redemption Project, as a place to focus on civics, rebuild non-partisan trust, and provide educational and emotional grace while learning about the news. I also have a column in Knox TN Today. My reporting and commentary have also appeared in other outlets including; Governing, The Arizona Capitol Times, South Florida Sun Sentinel, Police1, among other state and regional outlets.

Subscribe now

Behind the Reporting:

Why We Are Not Turning 190, 37,755 and 49,541 Into One Scary Statistic

Paid subscribers get the source-response ledger, data reconciliation, 2024 audit comparison and the questions needed to determine whether Tennessee has a statewide assignment backlog or a narrower Knoxville workload problem.

The strongest number is 190 — but its denominator is still missing

KPD gave TRP unusually specific language:

“Our current backlog of actionable tips awaiting assignment is 190 with the oldest tip being October 2025.”

Three elements are directly supported:

Read more

To some, bail is a way to keep dangerous people in jail.

To others, it is a system that punishes poverty.

To judges, bondsmen, defendants, victims, families and communities, it can mean something much more complicated.

In this Season 3 Systems Explained episode of The Redemption Project, Brandon Burley talks with Ken Good, an attorney from Texas who has spent decades working in and around the bail system.

Ken began his legal career representing doctors and hospitals, then developed a practice representing bondsmen and insurance companies across Texas. He has argued bail-related cases before the Texas Court of Criminal Appeals, the highest criminal court in Texas, and the Supreme Court of Texas. He has also written extensively on bail reform, criminal justice reform and what he believes works — and does not work — in the pretrial system.

The conversation begins with a basic question:

What is bail for?

Ken’s answer is grounded in the traditional purpose of bail: making sure a defendant returns to court while the case is pending.

That sounds simple.

But the details are where the system gets complicated.

Brandon and Ken discuss one of the central tensions in bail reform: whether people charged with similar crimes should be treated the same or whether judges should consider the individual facts of the person, the charge, the risk and the circumstances.

Brandon raises the example of two people charged with theft.

One person has a long criminal history.

The other has no record and stole groceries to feed a family.

Should they be treated the same?

Ken’s answer is no.

In fact, he pushes the point further. Even two first-time defendants charged with similar offenses may not present the same risk. One may have deep roots in the community, family nearby, stable housing and a job. Another may have no ties, no support and a much greater risk of failing to appear.

That is the heart of the episode.

Fair does not always mean identical.

Ken argues that the justice system has historically relied on individualized discretion because people, cases and risks are different. The reform movement, in his view, often pushes toward treating people the same based on categories, charge lists or automated tools.

But the criminal justice system does not deal with abstractions.

It deals with people.

And people are not identical.

The conversation moves into cashless bail, which Ken describes more broadly as simple release — releasing someone on a promise to return, personal recognizance, zero bail or other similar mechanisms depending on the state.

Ken is critical of these systems when they are used too broadly. He argues that simple release mechanisms often produce higher failure-to-appear rates than private surety bonds and can cause major problems for courts because cases cannot move forward until defendants return.

Brandon compares that to Tennessee’s citation-release system, where a person signs a promise to appear in court. Ken explains that in Texas, citation release is used much less often and generally for low-level offenses.

From there, the conversation turns to what happens when someone commits a new offense while already out on bond. Ken explains a new Texas law requiring that, when a person is out on bond and arrested again for a new felony in the same county, only the judge from the first case can set bail on the second case. He also says the statute requires judges to consider revoking or modifying the first bond.

That part of the conversation shows why bail is not just a release question.

It is a supervision question.

It is a court-management question.

It is a public-safety question.

The episode then moves into ankle monitors.

Ken does not argue that ankle monitors have no use at all. He says they can provide supplemental data and may have a role in high-risk cases.

But he rejects the idea that ankle monitors are a replacement for the bail industry.

His concern is practical. Monitors can generate false alerts, dead-battery notifications, exclusion-zone violations and other reports that desensitize the system. He also argues that defendants learn how to work around them, including cutting off a monitor late on a Friday to gain a head start before a warrant is issued.

Ken also raises the cost issue. If someone is paying daily fees for an ankle monitor over a case that lasts years, that person may end up paying far more than they would have paid through a bondsman.

That is an important Season 3 distinction.

A tool can sound like reform.

But the question is how it works in real life.

Who pays?

Who monitors?

Who responds?

How fast?

What happens when the alert is false?

What happens when it is real?

What happens on Friday afternoon?

What happens when the case lasts four years?

The conversation then moves to the role of bondsmen.

Ken argues that private surety bonds provide more than a financial transaction. He says bondsmen often provide reminders, supervision, family contact and a relationship that can help get someone back into court if they miss a date.

Brandon raises the example of a family member putting something at risk to help a loved one get out. Ken explains that the details vary by state, but he agrees with the broader point: family involvement can matter.

That is another theme running through the episode.

Systems are not only statutes and software.

They are relationships, incentives and accountability.

The conversation then gets into one of the most debated ideas in pretrial justice: risk assessments.

Ken is strongly critical of using risk assessments to replace judicial discretion in bail decisions. He argues that risk assessments may be better at predicting group trends than individual behavior, that they can produce racially distorted outcomes, and that they often make mistakes in the direction of recommending release.

He also raises a key legal concern: if a risk assessment does not consider the seriousness of the pending charge, but state law requires judges to consider the charge when setting bail, then the tool may not comply with the legal framework.

Brandon presses the point with a practical example:

If someone has no criminal history but is accused of shooting multiple people, should a tool treat that person as low risk simply because it is a first offense?

That question captures the problem.

The presumption of innocence matters.

But bail is not the trial.

Ken emphasizes that presumption of innocence applies to the trial itself. Bail is about assurance — what it takes to ensure that a person returns to court and, depending on the jurisdiction, whether there are public-safety concerns while the case is pending.

The episode also covers the political pendulum around criminal justice reform.

Brandon uses the image of Newton’s cradle: when policy swings too hard in one direction, the reaction may swing just as hard the other way. Ken agrees and argues that public safety eventually becomes the demand when crime rises and the public believes the system is failing.

That leads into Tennessee’s proposed constitutional amendment and the broader question of whether reform can swing too far toward release or too far toward detention.

Ken’s position is that the middle ground should involve accountability, private surety bonds and sparing use of simple release mechanisms. He worries that one extreme releases too many people without accountability, while the opposite extreme may hold too many people and create costs the public is not prepared to pay.

The conversation does not stay only on bail.

It also moves into jail overcrowding, incarceration costs and mental health.

Ken argues that part of the problem is that the United States has moved much of its mental-health burden into the criminal justice system. Brandon agrees from his experience in East Tennessee, where the closure of mental-health facilities meant law enforcement and jails became the default place for people who needed very different kinds of care.

That may be one of the most important parts of the episode.

The jail is not just holding people accused of crimes.

In many communities, it is also functioning as the largest mental-health facility.

That creates problems for defendants.

It creates problems for officers.

It creates problems for sheriffs.

It creates problems for courts.

And it creates problems for the people who needed treatment long before they ended up in a cell.

Ken shares a personal example involving his sister, who suffered a severe injury and later dealt with a long struggle involving prescription drugs. He describes the importance of family involvement and says that, in his view, families often provide the best chance for success for people with serious mental-health or substance-related problems.

But Brandon adds the necessary caveat:

Not every family has the means, resources or ability to provide that level of care.

That is the kind of tension this episode keeps returning to.

The easy answer is usually too easy.

Release everyone is too easy.

Hold everyone is too easy.

Trust a computer is too easy.

Blame judges is too easy.

Blame poverty is too easy.

Blame crime alone is too easy.

The real system is harder.

It has defendants, victims, judges, bondsmen, families, mental illness, addiction, money, public safety, constitutional rights, political pressure and imperfect tools.

Ken closes by pointing listeners to pbtx.com and thebailpost.com for more information about his group, their blog and their podcast.

This episode is a reminder that criminal justice policy cannot be reduced to a bumper sticker.

Bail is not only about money.

Ankle monitors are not only about technology.

Risk assessments are not only about data.

And reform is not automatically good just because it is called reform.

The better question is the Season 3 question:

How does the system actually work — and what happens when we change one part without understanding the rest?

I am a retired detective and criminal justice / government educator based in Tennessee. I founded The Redemption Project, as a place to focus on civics, rebuild non-partisan trust, and provide educational and emotional grace while learning about the news. I also have a column in Knox TN Today. My reporting and commentary have also appeared in other outlets including; Governing, The Arizona Capitol Times, South Florida Sun Sentinel, Police1, among other state and regional outlets.

Subscribe now

In Tennessee, a criminal fine can be only one part of the bill. Court costs, litigation taxes, later fees and collection charges can turn the amount announced in court into a much larger financial obligation.

By Brandon L. Burley
Systems Explained

The judge says:

“$200 fine.”

That sounds like the bill.

It may only be one line of it.

In Tennessee, a criminal ca…

Read more

The Redemption Project Newsroom

A stolen vehicle passes a roadside camera.

The license plate is compared with a law-enforcement hot list.

An alert reaches an officer, who confirms the plate, stops the vehicle and returns it to its owner.

In that case, the technology worked.

The vehicle was located. The owner recovered property. Police may have identified a suspect who otherwise would have remained unknown.

That is meaningful.

It also does not prove that the camera prevented a vehicle from being stolen.

Those two outcomes are often folded together when police departments, vendors and elected officials discuss automated license plate readers.

A camera may help police find a vehicle.

It may help investigators clear a case.

It may contribute to an arrest.

Whether installing a larger network causes crime to decline is a different question.

The evidence is considerably stronger for the first claim than the last.

That distinction sits at the center of the seventh article in this series.

Three claims that should not be treated as one

The first article in this series explained how Flock Safety’s automated license plate readers turn roadside observations into searchable records.

The system can produce several kinds of results.

It can generate activity: plate reads, searches and alerts.

It can produce investigative outputs: recovered vehicles, identified suspects, arrests and cases in which the data supplied a lead.

And it may contribute to broader public-safety outcomes: improved clearance rates, reduced theft or lower crime.

Those categories are connected.

They are not interchangeable.

A department may scan millions of plates and generate thousands of alerts without knowing how many alerts were accurate, how many produced stops or how many materially changed a case.

An officer may recover a stolen vehicle without preventing the next theft.

Investigators may use Flock while solving a homicide even though witness statements, phone records or physical evidence would have solved the case without it.

The narrow claim may still be true.

The system helped.

The broader claim — that the system caused the outcome — requires more evidence.

Flock says its technology assisted one in five cleared cases

Flock published its first annual Impact Census in May 2026.

The company surveyed nearly 700 law-enforcement customers across 43 states. Agencies that formally tracked Flock-related outcomes supplied counts. Agencies without formal tracking provided estimates of the percentage of cases in which Flock contributed.

Flock then matched responses with FBI crime and clearance data and weighted the findings according to agency size and crime volume.

The company estimates that its technology supported more than 1 million criminal investigations and incidents during 2025.

Read more

The party names survived civil rights, crime, trade, guns and cultural change. The coalitions underneath them did not.

By Brandon Burley
The Redemption Project Newsroom

Take the party labels away for a moment.

One party says illegal immigration must be stopped, celebrates a larger Border Patrol and promises tougher enforcement against people working in the…

Read more

Tennessee residents can pay substantially different fees for the same basic act: using a card to pay a government bill online. The government often says it does not keep the extra money. So who does — and what exactly is the public buying?

By Brandon L. Burley
Systems Explained

The tax bill is $5,000.

Pay it by credit card online in Knox County and the publ…

Read more

Tennessee records show that money sent by families can move through deposit fees, phone charges, commissary sales and, in some county jails, commissions that flow back to the facility. The person incarcerated uses the service. Someone outside often supplies the money.

By Brandon L. Burley
Systems Explained

The money might come from a mother.

A wife.

A grandf…

Read more

The Redemption Project Newsroom

Johnson City did not approve a few roadside cameras.

In July 2025, the city authorized a 10-year agreement with Flock Safety for 145 license plate readers, 145 pan-tilt-zoom video cameras, two mobile security trailers, gunshot-detection technology and software for a real-time crime center.

The first-year payment was $300,000. There is no year-two payment. The city’s published payment schedule calls for $970,375 annually during years three through 10.

That places the full agreement at more than $8 million.

Johnson City says installation is expected to be completed by Aug. 1, 2026.

The city also says its police department automatically shares Flock data with law-enforcement agencies within a 500-mile radius and has the ability to connect with more than 5,000 other agencies using the company’s platform.

That is more than a camera purchase.

It is a local government buying a place inside a regional surveillance network.

Johnson City has published considerably more information about its system than many communities. The city has disclosed the equipment, cost, general placement strategy, retention period, local sharing partners and portions of its use policy.

That transparency is useful.

It also reveals how many questions Tennessee does not require every agency to answer.

There is no complete statewide public inventory showing how many plate readers are operating, where they are located, which companies provide them, how local sharing settings are configured, how often officers search them or whether the searches produce arrests, prosecutions and measurable reductions in crime.

Tennessee regulates how long ordinary plate-reader data may be stored.

It does not require a full public accounting of what happens while the data exists.

The network was not built through one state plan

The first article in this series explained how automated license plate readers convert roadside observations into searchable records containing a vehicle image, plate number, time and location.

In Tennessee, that capability did not arrive through one statewide deployment.

Cities and counties bought systems separately.

Some began with a handful of cameras at major roads or city entrances. Others added cameras through annual budgets, police drug funds, grants, neighborhood partnerships or contracts with private businesses.

The result is not one centrally designed Tennessee network.

It is a collection of local systems that can become connected through vendor platforms, interagency agreements, regional intelligence centers and requests from other law-enforcement agencies.

By February 2023, WBIR reported that more than 1,000 cameras were already capturing license plates across Tennessee. That count included multiple agencies and was not limited to one vendor.

Deployment has continued since then.

Johnson City previously operated 11 license plate readers. Its new agreement increases that number to 145 while adding an equal number of general-purpose video cameras and several other surveillance and response tools.

Kingsport, Knoxville, Chattanooga, Memphis, Murfreesboro and numerous smaller cities and counties also use plate-reader or real-time crime center technology in various forms.

But no statewide dashboard allows a Tennessee resident to see the network as a whole.

Subscribe now

A Tennessee survey found rapid growth and uneven rules

A 2024 survey conducted by the University of Tennessee Municipal Technical Advisory Service and the Howard H. Baker Jr. School of Public Policy and Public Affairs offers one of the clearest available views.

Fifty-three Tennessee municipal police departments participated.

Two-thirds reported using automated license plate readers.

Read more

Uncashed checks, forgotten deposits and abandoned accounts can eventually become Tennessee unclaimed property. State law says Tennessee does not own that money. The same law generally sends the cash into the state general fund while Treasury remains responsible for paying valid owners who later appear.

Maybe you moved.

Maybe an insurance company mailed a …

Read more

You apply for the apartment, job or loan. But before anyone decides, another version of you may already be in the room — assembled from records, histories, scores and data you may never have seen.

You apply for an apartment.

Your income is stable.

You fill out the application.

Pay the fee.

Wait.

Then the answer comes back.

No.

Or maybe the apartment is still av…

Read more

In this Season 3 Systems Explained episode, Brandon Burley talks with Nick Nugent, a University of Tennessee law professor who teaches technology law and constitutional law, about automated license plate readers, public safety, privacy and the difference between what may be legal and what kind of surveillance society we actually want.

The conversation separates two questions:

Is this constitutional?

And even if it is lawful, is this the policy Tennessee communities want?

Brandon brings his law enforcement background into the discussion, explaining why he once helped set up a limited Flock system but now has serious concerns about mass vehicle-location tracking, private third-party data collection, weak audit controls and the ability to map a person’s movements over time.

This is not an anti-police episode.

It is not a pro-surveillance episode.

It is a civic explainer about technology, public safety, privacy and the rules that should govern powerful tools before the tools become normal.

I am a retired detective and criminal justice / government educator based in Tennessee. I am a commentary write for Tennessee Lookout and a weekly columnist with Knox TN Today. My work examines public policy, public safety systems and civic responsibility. My reporting and commentary have also appeared in Governing, The Arizona Capitol Times, South Florida Sun Sentinel, Police1, among other state and regional outlets.

Subscribe now

The Redemption Project Newsroom

The surveillance camera was mounted above an intersection in Bakersfield, California.

404 Media reporter Jason Koebler stood beneath it, pulled out his phone and opened a livestream.

He could see himself.

The camera was recording him from above while the same footage appeared on his screen without a username or password.

Anyone who found the exposed interface could watch too.

The camera was supposed to give police and public-safety officials a better view of the intersection.

Instead, it was streaming that view to the open internet.

Investigative reporting published in December 2025 found that livestreams connected to at least 60 Flock Safety Condor cameras were publicly accessible. Researchers reported finding cameras watching intersections, parking lots, walking trails, playgrounds and other public spaces across the country.

Flock said the problem resulted from a limited configuration issue affecting a small number of devices. The company said affected customers were notified and the exposure was corrected.

The incident did not expose Flock’s primary license plate reader database.

That distinction matters.

It does not make the exposure unimportant.

These were not Flock’s standard plate readers

Article one focused on Flock’s Falcon automated license plate readers.

Falcon cameras generally capture vehicles as they pass. They create searchable records based on plates, vehicle characteristics, time and location.

The cameras involved in the December exposure were different.

Flock’s Condor cameras provide live and recorded video. Some models can pan, tilt and zoom. According to the company’s product materials, the system also can use artificial intelligence to detect and follow people or vehicles moving through a camera’s field of view.

A Condor camera may watch an intersection, park, parking lot, trail or other public area continuously rather than waiting for a vehicle to cross a fixed capture point.

Flock markets the cameras as tools for situational awareness, officer safety and investigations. An operator can watch an unfolding emergency, review recorded video after a crime or follow movement across a monitored area.

Read more

The Redemption Project Newsroom

San Francisco police did not give federal or out-of-state agencies direct access to the city’s Flock license plate reader system.

The agencies did not hack the network.

They did not steal an officer’s password.

They did not exploit a software malfunction.

They still searched San Francisco’s data nearly 300 times

A police audit released in June 2026 found that analysts working through two regional information-sharing organizations had conducted 299 improper searches on behalf of federal and out-of-state agencies. The outside agencies included the Drug Enforcement Administration, Internal Revenue Service, Bureau of Alcohol, Tobacco, Firearms and Explosives, and state agencies in Oregon and Washington.

Those agencies were not authorized under California law to search San Francisco’s automated license plate reader data.

But a regional intelligence center was.

That distinction sits at the center of the third article in this series.

The most important question is not always whether an outside agency has direct access to a community’s cameras.

It is whether someone who does have access can search the cameras on that agency’s behalf.

Sharing is part of what makes the system useful

As the first article in this series explained, Flock license plate readers create searchable records showing when and where vehicles were observed.

A single police department can search its own cameras.

The system becomes more powerful when departments combine their reach.

A vehicle connected to a kidnapping may leave one city, cross several counties and enter another state. A burglary crew may strike multiple jurisdictions. A missing person may pass through a camera network hundreds of miles from home.

An agency limited to its own city boundaries may lose the vehicle.

A connected network may find it.

That is not an accidental feature of Flock’s system.

It is one of its primary selling points.

Flock says customers own the data collected by their cameras and control whether that data is shared. According to the company, an agency can limit searches to its own cameras, establish one-to-one sharing relationships, share within a geographic radius or state, or participate in a broader National Lookup network.

National Lookup allows participating agencies to search a complete, known license plate across cameras whose owners also have opted into the network.

That collaboration can be valuable.

It also means a camera purchased by one town may become part of an investigative network extending far beyond that town.

Subscribe now

There is more than one kind of access

Public discussions about license plate reader sharing often treat access as a simple yes-or-no question.

Does another agency have access?

The better answer may require several questions.

Read more

The Redemption Project Newsroom
Systems Explained Desk

A business earns a profit.

A city does not.

When residents ask whether Knoxville “made money” or whether the city is “profitable,” they are usually asking a different question: Is the city financially healthy? Can it pay its bills, maintain roads and public facilities, provide services…

Read more

The Redemption Project Newsroom

The car had been checked for a GPS tracker.

Nothing was found.

But according to a 2024 arrest affidavit, a Shelby County sheriff’s deputy still appeared to know where his former wife was going.

At one point, she received a message asking, “Where are you going?”

Investigators eventually found another possible explanation.

An audit showed that Deputy Thadius Gordon had searched the same Tennessee license plate approximately 20 times in four days and more than 100 times overall using a law-enforcement database, according to Action News 5. Gordon was charged with stalking and harassment and relieved of duty with pay. His attorney disputed the allegations and said the full circumstances had not been presented.

The database did not prevent the alleged conduct.

It documented it.

That distinction sits at the center of the second article in this series.

As the first article explained, automated license plate readers can turn individual roadside observations into a searchable history of where and when a vehicle was seen.

That history can help locate stolen vehicles, identify a suspect car or find a missing person.

It also can be searched by someone who already knows the plate number for reasons that have nothing to do with a criminal investigation.

At least 22 reported cases

The Institute for Justice, a public-interest law firm challenging automated license plate reader networks in several cities, has identified at least 22 reported cases in which officers allegedly used license plate reader data to monitor current partners, former partners, suspected romantic rivals or other people in whom they had a romantic interest.

The organization’s article was updated July 13, 2026.

Most of the reported incidents occurred since 2024. Nearly all of the officers were criminally charged and lost their jobs through termination or resignation, according to the organization’s review.

Only a few of the cases were initially discovered through internal investigations.

Most reportedly came to light after the person being monitored complained about the officer’s behavior.

Twenty-two is a serious number.

It also needs to be described precisely.

The count is not limited to Flock Safety. It includes Flock and other automated license plate reader systems.

It is not a count of every kind of license plate reader misuse. It excludes cases involving searches for political, immigration or other personal purposes when no romantic interest was documented. It also excludes misconduct that was never reported publicly, never detected or resolved internally without a public record.

Twenty-two is therefore not the total.

It is the documented floor for one specific category of reported abuse.

Subscribe now

The pattern is difficult to dismiss

The cases span officers, deputies, detectives, police chiefs and at least one elected sheriff.

The alleged search counts are not generally limited to one curious or mistaken lookup.

Read more

The firefighters have left.

The flames are out.

The homeowner is standing in front of a damaged house trying to answer the next question:

What happens now?

For generations, fire claims followed a familiar pattern. A fire damaged a home. An insurer evaluated the loss. Contractors repaired…

Read more

The Redemption Project Newsroom

A license plate is visible in public.

A searchable history of when and where a vehicle was seen raises a different question.

A police officer sitting beside a roadway has always been able to see a passing car, read its license plate and remember where it was headed.

That observation was limited by ordinary human constraints. The officer had to be there. The officer had to notice the vehicle. Unless something caused the information to be written down, the moment eventually disappeared.

Flock Safety’s automated license plate readers change that equation.

The camera does not simply see the vehicle. It creates a searchable record of the observation: the plate, the time, the location and visible characteristics of the vehicle. Those characteristics may include make, model, color, body type, damage, decals, roof racks or even the absence of a readable plate. Flock describes its product as searchable vehicle evidence, not ordinary video footage.

That distinction matters.

Police have always been able to observe vehicles traveling on public roads. What is new is that the observation no longer has to disappear when the officer looks away.

What a Flock camera does

This article focuses on Flock’s license plate reader system, commonly associated with its Falcon cameras. Flock also sells video cameras, gunshot-detection equipment, drones and other public-safety products. Those systems raise additional questions, but they should not be casually blended together.

A license plate reader is not the same as a general-purpose video camera.

When a vehicle passes a Flock license plate reader, the system captures an image and converts portions of that image into searchable data. An authorized user can search by a complete or partial plate number, but searches are not limited to a known full plate. Flock advertises search capabilities tied to vehicle type, color, make, model, time window, location and distinguishing features.

The system can also compare passing vehicles against law-enforcement hot lists. If a plate associated with a stolen vehicle, wanted person or missing or endangered person passes a camera, an alert can be sent to officers.

That alert is an investigative lead.

Officers still need to verify the plate, the alert status and the surrounding circumstances before taking enforcement action.

For investigators, the attraction is easy to understand.

A detective investigating a robbery may begin with a vague description: a dark SUV with a damaged rear bumper that entered the area shortly before the crime and left minutes later. Traditionally, the detective might spend hours collecting footage from homes, businesses and traffic systems. A searchable license plate reader network can narrow that work considerably.

It can help locate a stolen car, identify vehicles that repeatedly appeared near a series of crimes, find a missing person or corroborate whether a suspect vehicle was near a particular location.

Those are real investigative benefits. Pretending otherwise does not strengthen the privacy argument. It weakens it.

The better question is not whether the technology can help solve crimes.

It can.

The better question is what additional power is created when those observations are stored, searchable and shared.

Subscribe now

This is not just another traffic camera

A traffic-management camera usually watches a roadway or intersection so transportation officials can monitor congestion, crashes or road conditions. A real-time crime center may pull nearby camera feeds because a shooting, robbery or other event is happening now.

Read more

Tennessee’s fight over pharmacy benefit managers has moved from the legislature to federal court.

At the center of the dispute is the Fair Rx Act, a new Tennessee law aimed at separating pharmacy benefit managers, known as PBMs, from the pharmacies they own or control.

Supporters say the law is need…

Read more

Is this your thoughts about TDOT? You’re not alone.

TDOT’s public meetings on the I-40/I-75 corridor west of Knoxville are not just about traffic.

They are about how major road projects become real.

Before a bypass is built, before a new lane is added, before a bridge is replaced and before property…

Read more

One line from a 1941 book has stayed with me this week:

“We are only at the beginning…”

The line comes from The Redemption of Democracy by Hermann Rauschning, written during a time when democracy itself felt uncertain across much of the world.

Read more