Red light cameras and speed cameras have been praised and panned for taking the human element out of traffic enforcement. Now Ford is taking this one step further with a patent for a fully autonomous police car, filed in January.
According to details of the patent, the autonomous cop car would be capable of not only identifying and issuing citations for a passing car that’s breaking traffic laws, but also pursuing the alleged violator’s car if the AI processor deterimines that such an action is appropriate.
How exactly this decision would be made is left very vague in the documents, stating only that the processor would be “remotely executing one or more actions with respect to the first vehicle.” What this means is anyone’s guess. It could include turning on sirens and pursuing until the offending driver pulls over, or just snapping a photo of the driver’s license plate. Ford also seems to claim the cars could wirelessly communicate with the offending driver (despite no such technology existing yet) and let them know they’ve just received a ticket.
The patent also states claims that the car will be able to use a combination of sensors and data analysis to determine the best place to lie in wait for potential lawbreakers.
A patent filing does not guarantee that ticket-writing robo-cars will actually start patrolling the streets of course. But Ford does manufacture plenty of police cars. In fact, in 2015, 61 percent of new police vehicles purchases were Ford’s Police Interceptor SUV. But with so many companies working on and testing self-driving cars, it’s practically inevitable that they’ll hit the streets in some form. And if you can imagine a road full of driverless taxis moving people from place to place, it’s quite plausible that autonomous vehicles could begin showing up for other purposes, including law enforcement.
Removing all human elements from law enforcement raises many questions. While on the one hand machines can be prone to glitches, so are humans. But on the other hand and it might not be reasonable to expect an AI to occasionally (if not often) misunderstand often complicated laws. It also further complicates the issues already being faced by red light cameras. For example, every defendant in the U.S. has the right to cross examine their accuser. However, when one’s accuser is a machine, this is impractical if not impossible. Furthermore, if and when one such car writes up a ticket and a human tries to challenge that ticket, who will be believed? The person or the machine?