The Strategist

Death of Private Transport: What Will Future Cars Be Like?



01/18/2016 - 16:07



Autonomous vehicles have long been part of our reality. However, just as it happens with any mass innovation, the main obstacles to spread of unmanned vehicles is created not by physics, but politics.



Marc van der Chijs
Marc van der Chijs
Autopilots are not something radically new, not least because the aviation has been using it for decades, and the aviation industry is impossible without them now. Machinery, too, learned to operate without living drivers: a Google Car traveled through California and Nevada for more than 3 million kilometers. All automobile concerns - from Tesla to Toyota, Honda's up of Audi – are developing their own autopilots, not even hiding it. There has already been a year since the German company showed a car that drives without human help, and do it almost as fast as a professional racer.

Even transport companies cannot afford to ignore the trend. Uber’s CEO Travis Kalanick, whose company has doomed to the taxi services market to slow extinction, said recently that it would be funny if his own startup suffered the same fate. In a world of truly mobile machinery, Uber may lose value. So the company lured a few dozen experts in autonomous vehicles from Carnegie - Mellon University (Google Car’s creators) and is seriously considering replacing drivers with robots.

No one on the market has a doubt that by 2020, the technology of unmanned vehicles will work well enough to cancel human drivers behind the wheel. Yet, there is another problem. "Coming up with a new technology is not difficult, it is difficult to socialize it," - said three years ago the American top manager of Toyota John Hanson responsible for traffic safety and the environment.

Problems of socialization

Good news for fans is that the company is unlikely to refuse this particular technology. Road traffic accidents, according to WHO, are among the ten leading causes of death in the world, and the other nine are medical. If the number decreased by 10% because of the autopilot introduction, it would save 130 thousand lives a year and preserve a million people from severe injuries and disabilities. In fact, the decline may be more radical.

Google’s relatively small travel experience is encouraging – it is about 3 million km against 3 trillion kilometers, covered by all Americans on the roads of the United States during a year. In just a few years of testing, including on busy streets and roads, Google cars had just 12 minor accidents - each time no fault of their own. Judging by the accident, described in the company’s special report, they wouldn’t be even worth mentioning if happened with a human driver.

Another advantage of autonomous vehicles is that they are very profitable. Drivers and truckers are one of the most numerous professions, some major countries - US $ 1 billion or more to pay them. Imagine if robots were put in charge of long routes: fuel consumption would radically reduce with such a neat, never-sleeping, and non-temped by unofficial money made on the side. In the cities, automatic machines would not undercut each other and change lines just before the traffic lights – this could radically reduce number of traffic jams. Nevertheless, millions of people around the world will surely confront the economic and safety arguments. Taxi drivers and truckers are politically active and independent people who have something to lose, and, as the modern history shows, they know how to defend their rights.

Whatever it is, the drivers will lose and a new era will start. Coming out of autonomous vehicles will become expected, yet still a turning point. For the first time in human history, millions of consumer robots, able to kill or injure a person, will fill the streets of crowded cities of the earth. Programmers will develop applied rules governing behavior of autonomous vehicles, but everyone else will have to figure out how these rules should look like. The very need for such calculations puts humanity in the new circumstances. Many pragmatic issues and moral dilemmas that previously were ignored, will now have to be registered explicitly, no matter by law or program code.

Travel etiquette

Seat belt is an innovation that significantly improves safety of the passengers, but slightly reduces safety of pedestrians and other road users: drivers wearing seat belts are confident in their security, and drive a car in a relaxed manner. In the human world, the deal has a right to exist, is recognized as binding and has support of voters. In a world of robots, however, the balance will have to be calculated once again. A few lines of code will determine in which circumstances the machine will do all it can to save its passengers, and in which - passers. Hardly people will agree that a car’s owner, having paid to the manufacturer, will have the weather.

In October, a group of researchers from the Toulouse School of Economics and MIT published the first experimental work on the ethics of autonomous vehicles. Living people are asked to make a choice for a robot in various hypothetical situations involving risk of death in a traffic accident. For example, there was a choice between hitting a pedestrian on the lane or crashing into the fence and kill itself. The experiment showed that in France and America, people were shocked by the fact a robot decides the fate, at the time minimizing the overall loss. If total number of traffic accidents drops by several times thanks to the autonomous transport (and so does the car insurance’s price), then the robotics laws problem will be only philosophical.

A more difficult problem is that, like any robot, an autonomous vehicle is a dedicated computer connected to the Internet. Hence, it can be manipulated. This is not a hypothetical situation: at the beginning of the year two programmers, Charlie Miller and Chris Valasek, hacked a jeep and took the control right on the go (a journalist of Wired sitting behind the wheel suddenly lost the ability to slow down). They did it to attract attention to the problem of security of modern electronic systems. A few weeks after their achievement became known, Uber hired two programmers that they were responsible for the safety of the company's machines.

If a car is guarded from hackers, then will the vehicle’s owner himself be able to operate the car? Man has the right to buy an iPhone, and to jailbreak it to store his own files and run any own programs, not included in the official application store. But can he do the same with his car? Manufacturers of machinery will embedded in their cars a variety of programmatic constraints like speed limit or inability to suddenly change lines (moving across). Rather, even the traffic rules themselves will be set by software - the car, not the driver, would take the driver’s test. Governments in developed countries will surely ban private experiments with the dangerous robots once they are released on the roads. However, there still will be a competition: among corporations that develop software driving, and among a variety of jurisdictions with different driving regulations (the problem of left- and right-hand traffic will disappears in a moment).

Death of private transport

There are two approaches to designing and building autopilot machines. The first is preferred by automobile concerns: most of the time, driving occurs on autopilot, but if desired, the driver can take it on himself. Alphabet (Google’s enterprise) is trying to implement the second. Google cars are equipped with rudders only for test purposes; in the future, they will become full-fledged robot, devoid of manual controls.

If Alphabet succeed, it will become a fetcher to ban driving the human drivers - those who were to blame for all Google Car’s failures. At least some municipalities will want to ban human driving on its territory. And somewhere, the right to drive a car would be like in the era before Henry Ford's: the driver will have to pay high tax or a premium insurance.

A pragmatic approach tells to deprive people of incentives not only to drive cars, but also own them. An average American car (around 40% of world sales) is moved along the roads for about an hour a day, and stays idle the remaining 95% of the time. For practical engineers and managers involved in the development of autonomous vehicles, such a degree of inefficiency is an absolute scandal. For Uber and its clones, it’s a niche market worth tens of billions: taxi services are located at the intersection of cars and public transport.

Morgan Stanley analysts raised the forecast for Tesla shares by half, when suspected that the company could become a major player in the market of sharing autonomous vehicles. This is good news for the traffic security: nobody will tighten up speed limit of a car co-owned.