Ford's Robot Police Car Is No RoboCop

A view from a cyborg police officer in the 2014 film "RoboCop." Credit: Sony Pictures Before the RoboCop future arrives, a robot police car that pulls over speeding vehicles and issues tickets or warnings on its own could someday help ease a shortage of human officers at police departments across the United States. But the vision of a self-driving police vehicle described in a Ford patent also raises many questions about whether such technology is the right tool for law enforcement. The basic Ford patent description makes clear that this self-driving police car cannot do the job of a flesh-and-blood police officer. Instead, it focuses on the idea that "routine police tasks, such as issuing tickets for speeding or failure to stop at a stop sign, can be automated." It goes on to describe the self-driving police car's capabilities as follows: detecting traffic law violations by another vehicle, tracking and chasing the perpetrator vehicle until the latter pulls over, automatically checking the vehicle license plate or relevant driver's license with a central database, and deciding whether to issue a ticket or warning based on the violation. This idea sounds OK in theory: automating some of the dull and more routine tasks to free up human police officers for more urgent duties. That could help address growing shortfalls of police officers at law enforcement agenices in both rural communities and big cities such as Atlanta, Baltimore and Dallas; Houston's police department in particular has reportedly been short of as many as 2,000 officers needed to serve the country's fourth largest city. Yet in many ways, the patent description of the robot police car makes it sound like a solution in search of a problem. Go Go Traffic Law Enforcer Ford's patent describes a robot police car that is basically a mobile version of an automated red light camera or speed camera, rather than a RoboCop intended to chase down uncooperative suspects. The robot police car may still serve some broader law enforcement purpose as a deterrent—especially for human drivers who may tend to be more cautious in sight of a police vehicle—but it does not possess anything like the full capabilities of a human police officer. Ford's vision for an autonomous police vehicle leans heavily on the idea that many future cars—whether manually driven or fully driverless—will have the technology to wirelessly communicate with the robot police car. The approved patent specifically covers the concept of wirelessly receiving driver's license information from another vehicle. It also covers the concept of wirelessly transmitting messages regarding the robot police car's decision to another vehicle: a traffic ticket with a fine, a warning without a fine, or a message saying the vehicle is free to leave the scene. The Ford patent does allow for the possibility of the robot police car scanning a driver's license that a human driver holds out the window of the vehicle. But otherwise the robot police car in question seems best designed for pulling over and ticketing very cooperative vehicles or drivers, given that there is no mention of any contingency actions in case the suspect vehicle takes off and refuses to stop. No Such Thing as Free Policing One crucial question is whether local and state law enforcement agencies would actually consider this autonomous police vehicle a worthwhile investment in comparison with other technologies. The Ford robot police car or similar autonomous police vehicles would be much less attractive if their cost was significantly higher than those of potentially cheaper alternatives, especially given how local and state law enforcement agencies have already been struggling to recruit and retain human police officers due to budget issues. For example, cities and police departments have already been using automated enforcement technologies such as red light and speed cameras to detect traffic violations, sometimes as part of an automated system that can also mail tickets to homes of the perpetrators in question. Installing automated red light camera systems can cost between $67,000 and $80,000 per intersection. Separately, speed cameras installed on cars operated by the New York City Department of Transportation cost approximately $90,000 to $115,000, which doesn't include the vehicle costs. By comparison, the first generation of fully autonomous self-driving cars could easily cost hundreds of thousands of dollars per robot car. That's without even accounting for the additional cost of the robot police car's automated traffic enforcement system as envisioned by Ford, which would likely be at least as expensive as the fixed-installation red light cameras and speed cameras already in use today. So it's probably a reasonable assumption that Ford's robot police car would be more expensive than any standalone versions of automated camera systems. Where We're Going, We Don't Need Roads It's still possible that the robot police car's mobility would make it valuable as a flexible tool for law enforcement. Having a squad of autonomous police vehicles could theoretically be more cost-effective than, say, installing enforcement traffic light cameras at every intersection or main stretch of highway—especially if patterns of traffic violations continue to change among different geographical locations over time. But if a major U.S. city really wants a mobile technology that can automatically spot and ticket vehicles violating traffic laws, a robot police car may ultimately pale in comparison with flying drones. U.S. law enforcement agencies already operate surveillance aircraft and helicopters that can capture clear images of vehicle license plates--or even human faces--from up to a thousand feet away. Tomorrow's police drones could likely use machine learning AI to automatically spot many more traffic law violators from the air than a robot police car stuck on the ground. A robot police car may ultimately cost less than a long-endurance surveillance drone carrying the same equipment. But the utility of a drone or even a crewed surveillance aircraft could make the flying option more cost-effective than a robot police car in the long run. Maybe Not the RoboCop We Need Last but not least, the patent specifically mentions the capability to determine whether a vehicle is driverless or being operated by a human driver. That raises two related questions: Should we expect many autonomous vehicles in the future to actually break traffic laws by speeding or running red lights? And if not, what need is there for a robot police car that works best in pulling over cooperative smart vehicles? One of the big selling points for the supposed future of self-driving and fully driverless vehicles is that they would presumably reduce the number of driving accidents and traffic violations by gradually removing human drivers from the equation. Waymo and other companies developing driverless cars have repeatedly emphasized the idea that robot cars will drive safely and presumably in a law-abiding manner. At their dreamiest, self-driving car advocates talk about a future filled with autonomous vehicles that communicate with one another and a city's central traffic system in order to transform traffic gridlock into a beautifully choreographed dance of commuting efficiency. If that future system works as promised, a robot police car would seem fairly redundant in terms of policing autonomous vehicles. It's far more likely that driverless vehicles which might occasionally commit a traffic violation through malfunction could probably police themselves by self-reporting any problems to law enforcement and their human engineers overseeing the overall system. A robot police car might prove far more useful helping human police officers lay down the law to human drivers and people in general. But policing potentially uncooperative people rather than cooperative smart vehicles also presents a far greater challenge involving much more complex situations. It would undoubtedly raise the risks of something going terribly wrong, given the hypothetical scenarios of a robot police car trying to take out a suspect vehicle during a high-speed pursuit or somehow apprehending suspects fleeing on foot. Ford's idea could still find some use if robot police car capabilities evolve beyond those described in the patent. But a robot police car focused on policing law-abiding autonomous vehicles is unlikely to be the solution that cash-strapped police departments need right now, if ever.