Five things are reasonably well documented about these systems: what they collect, how often the alerts are wrong, who can search the data, the vendor's security history, and what the crime statistics show. None of it concerns the conduct of any local department.
An ALPR photographs every vehicle that passes, not only vehicles of interest. Each read records the plate, the time, and the location. Flock's system also indexes make, model, color, and distinguishing features such as bumper stickers, roof racks, and damage, which means a vehicle can be located without its plate being read at all.
Reads are retained whether or not they match anything. Historically the default window was 30 days; Flock lowered its default to 7 days in August 2026, though customers can configure it. Over that window the accumulated reads describe a vehicle's movement pattern rather than a single sighting, which is the basis of most legal objections to the technology. Courts have treated sustained location tracking differently from one-time observation in public.
Hit rates published by three local agencies range from about 7% to about 22% of reads. A hit means a plate matched a list, not that a crime occurred. None of the local portals publish how many hits were incorrect or how many produced a stop.
Where an outside body has counted, the figures are substantial. An audit by the Los Angeles Police Department's Office of the Inspector General reviewed 498 stolen-vehicle alerts over a two-month period: 337 correctly identified a stolen vehicle and 161 did not, a false-positive rate of 32.3%. LAPD let its Flock contract expire in July 2026.
The audit attributed most false alerts to the hotlist data rather than to camera misreads: recovered vehicles never removed from lists, outdated theft reports, and uncorrected data-entry errors. That is a records-maintenance problem in the source databases, distinct from the accuracy of the cameras themselves, and it is not detectable without an audit.
Separately, the Institute for Justice has catalogued stops in which a misread plate led officers to detain drivers who had committed no offense, in a majority of cases with firearms drawn.
The relevant local question is narrow: has any Hendricks County agency measured its own false-positive rate, and is the figure available?
Use policies generally classify ALPR data as sensitive and exempt it from public inspection. The same data is shared into a national network. Plainfield's transparency portal, the only local one publishing its sharing lists, shows access extending to agencies in more than 30 states and to several federal entities. Searches run under the rules of the querying agency, not the collecting one.
Misuse is documented at scale. A Washington Post investigation published August 3, 2026 identified at least 50 law enforcement officers charged with or accused of using plate reader systems for unauthorized surveillance of women they knew personally; Flock's system was involved in 46 of those cases. In 26 of them, investigators concluded the targets were partners, former partners, or women the officer wanted to meet.
How these cases surfaced is as relevant as the count. The majority came to light through victims, journalists, or outside review rather than through routine internal auditing. In Milwaukee, a victim identified the searches by looking up her own plate on a public log aggregator.
Indiana, August 2026
An Indianapolis Metropolitan Police Department officer was found to have run hundreds of ALPR searches concentrated on a small number of plates, with no legitimate law enforcement purpose. IMPD removed his police powers and system access on August 6, and Chief Tanya Terry suspended him with a recommendation for termination. Administrative and criminal investigations followed, and the department adopted a policy requiring searches to be tied to a case number.
The department's response was prompt and public. Reporting indicates the searches came to attention around the time a newspaper contacted IMPD rather than through scheduled auditing. The Indianapolis Fraternal Order of Police subsequently called for additional safeguards from the legislature.
These failures sit outside the control of any contracting agency. A town can write a strong use policy and still be exposed by a vendor-side lapse.
Atlanta has more Flock cameras than any other American city, over 5,000, and adopted the technology early. The Atlanta Community Press Collective compiled the police department's own figures from the FBI's Crime Data Explorer for 2021 through 2025, the period covering the buildout.
| Offense | 2021 clearance | 2025 clearance |
|---|---|---|
| Homicide | 53.4% | 48.0% |
| Rape | 37.7% | 37.7% |
| Motor vehicle theft | 9.9% | 10.2% |
Clearance rates also fell over the period for robbery, larceny, and shoplifting. 2025 was Atlanta's weakest homicide clearance year of the five.
Causation runs in several directions here and the decline should not be attributed to ALPRs. Departments nationally faced staffing shortages and a documented surge in vehicle thefts over the same years. The narrower observation is what it is: the most heavily surveilled city in the country shows no measurable clearance improvement across the period of its buildout, which is difficult to reconcile with claims that this scale of deployment produces large public safety returns.
Indiana has enacted no statute governing ALPR retention, sharing, or oversight, so these are decisions each department or council makes on its own. State legislators have indicated the subject is likely to come up in the next session.
Accurate, but the collection is consented to and can be switched off or left behind. Displaying a license plate is legally required.
Documented and not in dispute. It addresses a different question than what rules govern the data, since none of the requests above would prevent investigative use.
This is why the objection concerns the retained database rather than the act of observation. In the documented misuse cases, the people tracked were doing nothing private.
Consistent with the record. No payment occurs in any of the sharing documented here; access is granted through the network rather than sold.