My grandchildren won't need a driver's license

November 7, 2012

I predict that in 50 years, less that 1 car in 50 will be driven by a human. Robots are going to take that job, either as part of the car, or as a separate appliance that works with the car.

Here’s why: robots are going to be better, safer drivers than people. They don’t blink, they don’t drink, they don’t need to sleep, and they never panic (OK, so maybe they do sometimes). At some point, insurance companies are going to pick up on this, and reduce premiums for operators of self-driving cars. Over time, more people will opt for an autonomous vehicle on the basis of safety, particularly parents buying cars for their kids. Eventually that bonus will turn into a penalty paid by those who insist on driving themselves. In a generation, it just won’t seem worthwhile.

Computers are also much better than humans at optimising a car’s fuel efficiency. They watch that dial all the time, and they have the inputs and number crunching ability to figure out the most fuel-efficient way to drive, and when your electric car’s range is limited, that becomes really important.

This doesn’t mean that human-operated cars are going away completely. Driving started out as a sport, and it will go back to those roots.

There’s an interesting legal wrinkle, though, and it goes something like this: if a robot-driven car is involved in a collision, who is the responsible operator? Is it the human in the front passenger seat (perhaps the one behind the vestigial steering wheel)? Is it the manufacturer? Is it the last mechanic to service the car? Is it the operator responsible for maintaining the roadside network that gave the car its traffic data? Now imagine it’s a taxi, or a 50-tonne truck. That’s the question that could throw this entire prediction on its head. But it would be a real pity to lose all this potential just because somebody needed somebody to sue.

Asimov had an interesting thought on this matter (one of his many interesting thoughts, particularly on robots). In his stories, all robots were made and owned by the US Robotics Corporation (think Apple, but with more moving parts and an absolute monopoly), and leased out to users. This put the onus on USR to maintain and repair the robots, and gave them clear responsibility for the robots’ actions. Open source fans (myself included) would probably be quite resentful if this kind of arrangement was what we ended up with, but I dare say it (or something like it) is one of the more workable strategies for that particular legal issue.

Now that I think of it, taxi drivers, truckies and couriers are likely to be absolutely livid about robotic vehicles. I predict that unions will fight as hard as they can to keep this technology down. Because what we’re talking about is turning “driver” into an unskilled job. They’d sit behind the nominal controls and maintain the legal fiction that an individual is responsible for what the car is doing. But legal fictions are solemn things: if a law saying “all cars in operation must have a single human driver” came (or has come!) into force, then the entire question of robot cars becomes academic. As it is, an awful lot of vehicle-related law relies on the assumption that a human is in charge. It will take a lot of persuading to get a lawmaking body to give up an assumption of that magnitude.

So I’m going to stick to my original prediction, because I’m optimistic. I’m hoping really hard that it comes true.