Welcome back to Front Line Friday. This week is an editorial, and the topic is the wave of drones, license plate readers, and other technology moving into patrol work, and what actually changes officer capability versus what just moves money around. Front Line Friday is brought to you by Dead Air Silencers, whose support keeps this column going every week.
The sales pitch says technology is a force multiplier. The street says a tool only multiplies force if an officer can actually use it on the call where it matters, with the training time and staffing you actually have. Those are not the same claim. Most agencies buy the first one and get the second one, and nobody writes that part down.
This piece is not a product roundup of specific drone models or plate reader vendors. It is not a privacy-law explainer, and it is not an argument for or against surveillance policy. It is an operational read on three technologies entering patrol work right now, where each one actually helps, where it fails, and who it fits. The question is capability, not catalog specs.
Front Line Friday @ TFB:
Start with the failure mode, because it explains almost everything that goes wrong with tech on patrol. The failure mode is buying the asset and not buying the program. A drone can be a $12,000 line item. A drone program is a recurring cost in training hours, battery replacement, software licensing, data storage, FAA compliance, and at least one person whose job it partly becomes to keep the thing airworthy and current. Agencies routinely fund the first and starve the second.
This is not a story about lazy administrators. It is a story about structural incentives. A capital purchase is visible; it photographs well for the city council, and it draws down a grant before the grant expires. The recurring program cost is invisible; it competes with overtime and vehicles in the general fund, and it never gets a ribbon-cutting. So the incentive structure rewards buying hardware and punishes sustaining it. The predictable result is a closet full of technology that works for about six months and then quietly stops being current.
I have watched this cycle run its course more than once. The grant pays for the equipment and an initial training block. Year two, the certified operators transfer, promote, or leave. Year three, the batteries have aged out and the software subscription lapsed because nobody budgeted the renewal. The equipment is technically owned by the agency and operationally owned by nobody. That is not a capability. That is a depreciating asset that files its own incident reports to the storage room.
The useful question at the point of purchase is not "can we afford this." It is "can we afford this every year for ten years, including the labor to keep people current." If the honest answer is no, the capability does not exist regardless of what sits in the case. The operational consequence is that an officer reaches for a tool mid-call and finds a dead battery and a lapsed certification, which is worse than never having bought it, because now the plan assumed it would work.
Small uncrewed aircraft are the clearest case of genuine capability gain in recent patrol technology, and also the clearest case of hidden overhead. Used well, a drone gives you an overhead view of a scene in under two minutes that would otherwise require a helicopter you do not have or a foot search you cannot staff. On a barricaded subject, a missing elderly person in rough terrain, or a building search before entry, that view changes decisions and can keep people out of harm's way. That part is not hype.
Here is where it fails. A drone is only as fast as the certified pilot who can legally fly it, and certification is not a one-time event. Under current FAA rules, agency operators generally fly under a Certificate of Authorization or a Part 107 remote pilot certificate, and Part 107 requires a recurrent knowledge check every 24 months. The short version: the license is not a wall decoration, it is a sustainment cost with a clock on it. Night operations and flights over people carry their own additional requirements. If your pilot pool is three people and two of them work days, your 0300 barricade does not have a drone no matter what the policy binder says.
The second failure mode is the gap between what the rating tests and what the mission demands. A Part 107 certificate tests airspace knowledge, weather, and regulations. It does not test flying a thermal camera through a treeline in wind at night looking for a hypothermic subject. Those are different skills, and the second one decays fast without reps. An agency that counts current certificates and assumes current capability is measuring the wrong thing. The operational consequence is a pilot who is legal to fly and not actually good at the one flight that matters, discovered in real time on the worst possible call.
Who does a drone fit? Agencies with enough call volume and enough staffing to keep a real pilot rotation current, and a clear doctrine for what the drone is for. The Drone as First Responder model, where an aircraft launches from a fixed site to a call ahead of officers, is a genuine capability for dense jurisdictions with the infrastructure to support it. For a rural agency with six deputies covering 800 square miles, a single pilot and one aircraft can still be worth it for searches, but only if someone owns the sustainment honestly. The aircraft is the cheap part. The current, competent pilot is the expensive part, and the expensive part is the actual capability.
Automated license plate readers are the technology most likely to be oversold and misunderstood in the same breath. A fixed or mobile LPR scans plates, runs them against hot lists, and alerts on hits. On stolen vehicles, Amber and Silver alerts, and wanted-vehicle investigations, that is a legitimate and sometimes decisive tool. It does work that no officer scanning plates by eye at 45 mph can match in volume.
Now the failure mode, and it is a serious one. An LPR hit is a lead, not a fact. The documented error cases are not rare edge cases: misreads of a character, a plate that was stolen or reassigned, a hot-list entry that was never cleared after the vehicle was recovered. There have been real incidents where officers conducted high-risk stops on innocent drivers, including children and families at gunpoint, because the system alerted and the humans treated the alert as confirmed truth instead of as a reason to verify. The technology did not fail there. The procedure around it failed, because the alert was trusted past what it actually establishes.
The distinction that has to be trained into muscle memory: the reader tells you to look, it does not tell you what you found. Before you act on an LPR alert, the plate gets confirmed visually, the hot-list entry gets confirmed as still active, and the stop gets built on the officer's own verified observation. That is not bureaucratic drag. It is the single step that separates a good stolen-vehicle recovery from a civil suit and a family's worst night. The operational consequence of skipping it is that your agency inherits the system's error as your use of force.
Who does an LPR fit? Nearly any agency can get value from it on auto theft and wanted-vehicle work, because that is squarely what it does. The question is not whether to use it but whether your policy and your field training force verification before action, every time, without exception for the hit that looks obvious. The agencies that get burned are the ones that bought the camera and never built the procedure. Technology plus no procedure is not capability. It is liability with a subscription fee.
Body-worn cameras are old news as technology and still misunderstood as a system. The common framing treats the camera as a surveillance tool pointed at the officer. The more useful framing, and the one I hold, is that thorough documentation is a tool of officer protection and genuine accountability at the same time. Those are not in tension. The same footage that clears a good officer convicts a bad one, and both outcomes serve the public and the profession.
The failure mode here is not the camera. It is the data pipeline behind it. A body camera program generates enormous volumes of footage that must be tagged, categorized, retained on schedule, and produced on request. The recurring cost is storage and the labor to manage it, and that cost scales with every camera you add and every retention requirement in your policy. Agencies that bought cameras on a grant and never funded the storage and the records staff end up with footage they cannot find, cannot produce in time, or delete by accident. The camera captured the truth. The system lost it. That is worse than no footage, because now the gap looks like concealment.
There is a second, quieter failure: activation discipline. A camera that is not recording during the thirty seconds that matter protects nobody. Reliable activation is a training and habit problem, not a hardware problem, and it degrades under exactly the stress where the footage matters most. The fix is not a better camera. It is drilling activation until it happens without conscious thought, the same way you drill a draw. The operational consequence of weak activation discipline is that your best protection against a false complaint is dark during the one call where you needed it.
Who does a body camera program fit? Every agency that can fund the back end, which is most of them if they count the real cost up front. The mistake is treating the camera as the purchase. The purchase is a records system with a camera attached to it. Buy it that way and it is a genuine asset. Buy it as hardware and it becomes a liability generator the first time a defense attorney asks for footage you cannot locate.
There is a thread running through all three of these tools that does not show up on any spec sheet, and it is the one that should worry you most. Technology that reduces workload also reduces attention, and a tired officer leaning on an automated alert is exactly the officer most likely to accept that alert without checking it. This is where tech and fatigue stop being separate problems.
Fatigue is a physiological safety issue, not a character issue, and the research on it is blunt. Sustained wakefulness degrades performance the same way alcohol does. After roughly 17 hours awake, cognitive and reaction performance is comparable to a blood alcohol concentration of about 0.05 percent. Push to around 24 hours awake and the degradation is comparable to roughly 0.10 percent, which is over the legal driving limit in every state. The short version: the back half of a double shift leaves an officer impaired by a standard we would arrest a driver for. That officer is now being handed a system that says "this plate is stolen, go."
Put those two facts together. An impaired-by-fatigue officer plus an automated alert that invites compliance is a predictable path to the error cases I described earlier. The technology did not create the fatigue, but it changes what the fatigue costs, because it offers a confident-looking answer at the exact moment the officer has the least capacity to question it. This is automation bias meeting shift work, and it is structural, not personal. The operational consequence is that your newest capability is most dangerous on your most understaffed, most over-worked shift, which is also the shift where you deployed it to cover the gap.
The fix is not to ban the tools. It is to design the verification step so it survives a tired brain, which means making it a hard procedural stop rather than a judgment call. A judgment call fails under fatigue. A checklist item that cannot be skipped fails less. That principle, build the safeguard into the process so it does not depend on the officer being sharp, is the whole game with patrol technology.
One more structural point, because it decides whether any of this scales. Every new tool adds cognitive load, and cognitive load compounds across a shift. Three drone platforms, two camera systems from two grant cycles, and an LPR interface that updates its menus quarterly is not a technology program. It is a pile of unrelated interfaces an officer has to context-switch between under stress, and context-switching under stress is where mistakes live.
Standardization is a force multiplier precisely because it reduces that load. One camera platform fleet-wide, one drone type, one LPR interface, means a transferred officer is current on day one and a mutual-aid partner can pick up your controller without a tutorial. The cost of standardization is giving up the ability to chase the best-of-breed product in every category. The payoff is that officers can actually use what you bought. For a capability that lives or dies on whether a stressed human can operate it correctly, that trade is almost always worth it.
This matters most in the career-long view. An officer who learns one standardized tool set builds deep competence across years instead of relearning interfaces every grant cycle. The agency that chases novelty resets its people to novice every few years and calls the churn "modernization." The decision to standardize early compounds the same way a good training habit compounds: quietly, across a whole career, into capability you cannot buy with a single purchase order.
Fire and EMS are living the same story with different acronyms. Drones for scene size-up on structure fires and wildland incidents are a real and growing capability, with the same sustainment trap: the aircraft is cheap, the current pilot is not. Thermal imaging has already proven that a tool only helps if the user has enough reps to interpret what the screen is showing, because a thermal image of a smoke-filled room is not self-explaining under stress.
EMS has its own automation-bias problem in the cardiac monitor that suggests a rhythm interpretation. A fatigued medic on the back half of a 24-hour shift leaning on the machine's read instead of confirming it is the same trust trap, with a patient instead of a plate. And fire and EMS run even longer shifts than most patrol, so the fatigue math is worse, not better. The lesson transfers cleanly: fund the sustainment, drill the verification until it survives exhaustion, and standardize the interface so a tired provider is not fighting the tool and the emergency at once.
That's Front Line Friday for this week: patrol technology multiplies capability only when you fund the sustainment, build verification into the process so it survives a tired officer, and standardize hard enough that people can actually use what you bought. The hardware is the cheap part and the easy part. The program is the whole point, and the program is what procurement keeps forgetting to buy.