A young engineer was speaking to me about the future of cars and roads. The addition of artificial intelligence to cars is ongoing, and will soon reach the stage, he says, where it will become clear that cars in certain urban areas will not be allowed to drive with human drivers at the wheel.
Such an outcome assumes a great deal of progress in resolving a host of issues, technical, social and political.
The implications of increased intelligence in cars – up to the point where humans can be replaced as drivers – go on and on.
- ownership versus renting
If cars can be rented by the hour or by the occasion, the incentives to own a car may go down. Cars usually sit in the driveway or the parking lot for most of the day. Imagine that cars are basically taxis, and that the ownership (whoever they or it may be) cleans, maintains and provides cars on much the same basis as taxis, but with no taxi driver. You would summon a car as you would an Uber taxi, and it would show up at your location, but without the driver. Step in and the car will drive you to your destination.
- traffic signals
Your community is strewn with stop signs, lights, and painting of signals on the road. Imagine that the driving rules for every intersection are communicated by local networks to the cars within reach of the signal, and that cars communicate by networks to each other in constant Bluetooth-style to adjust momentum (direction and speed). Once cars are self-directing, if the destination has been selected by the passenger, then a huge infrastructure of visual signs would be replaced with an electronic infrastructure. As a pedestrian, you may need a sign as to where you can cross, but the governing software of cars will ensure that, within the limits of the laws of physics, cars will not be able to hit you.
- legal compulsion
It will be argued that the full benefits of the driverless automobile system will only be realized when people are legally obliged to switch over from the human driver to local network control. The law will compel drivers in certain areas to surrender control, and in all likelihood the car will simply adjust by becoming integrated with the local network, on the supposition that there is a private automobile entering the local network space.
The sign saying “you are now entering Such and Such” municipality also acts as the point where the car – not your car but “the” car – passes from the control of one network to another, just as a cell-phone call is passed from one tower to another. The car in which you are riding has become a physical instantiation of a telephone call.
The consequences of this driverless system are expected to be:
1) drastic reduction in the amount of society’s resources dedicated to automobiles, as the use of each car intensifies. This may mean fewer cars, or less social investment in related automobile technologies, or lowered energy consumption. It may allow for quicker transitions to newer propulsion technologies.
2) legal liability will be need to be worked out between the software makers (General Motors, Toyota, Apple whoever) that make the car control software, the cities which install the driverless networks, and insurance companies for both sides.
Some of the negative effects will be:
1) loss of autonomy and privacy, but as computer technology invades everything, the loss of autonomy will long precede the transfer to the automated driverless system spoken of here. You are already being followed by your GPS and other technologies in your car, even if you still drive it. Mandatory guidance systems will not change the trackability of cars.
2) Every car will become like a taxi. The cleanliness, appearance, and maintenance level of your car will depend on the previous occupants, and on which company owns them, and some companies will be better than others. Given the human propensity for status distinctions, people will pay for better cars by belonging to better car-cooperatives.
Cultural and social resistance will take a long time to be overcome.
First, the software to run all this must work seamlessly and efficiently to figure out the dozens of social and safety rules that govern human transactions in every driving situation. Consider four-way stops which can be a ballet of mutual recognition. The mutual interchange of signals among cars and the successors to stop signs and traffic lights must work out in a faultless protocol. WIll drivers be allowed to assume control, and in what circumstances?
This leads to the second huge issue: trust. It is likely that failures will become as rare and nearly as deadly as airplane accidents. Imagine a breakdown of signals, or the failure of protocols, on a highway where hundreds of cars are hurtling on autopilot. It will take a long while before people can trust the state of the system to be sufficiently faultless that getting into a car is as safe as getting into an airplane.
Inconvenience is the third major reason for resisting. Private ownership of cars may be as irrational as the private ownership of power tools, from the perspective of efficiency of use, but people do not like systems of common or collective ownership for good reason. Some people are slobs, others neatfreaks. Some use their cars as mobile filing cabinets. So private ownership will likely continue, even in the brave new world of automated driverless cars. Thus the argument for the driverless car system is not an argument for the abandonment of private ownership, but it will increasingly make private ownership look as anachronistic as a CD or record collection.