The latest model of Tesla can drive pretty much by itself on motorways, but it still requires a human being operating the steering wheel and pedals. It has multiple sensors analysing huge quantities of information from its surrounding and yet, it still requires John Smith to be attentive at what it does. If the technology is here, why are we still driving cars?

One of the first issues that comes to mind is reliability. Let’s imagine that vehicles and their technology can be incredibly reliable. Close to engineering perfection, but obviously not 100%. If you have had any kind of electronic or mechanical devices in your life--if you read this there is a big chance you do--you already know that this level of reliability is impossible. How many times did you have a problem with your smartphone, with your TV, with your computer, or with your car? Surely quite a few times. So, would you really want to be a passenger of some sort of transportation that is not fully reliable?

“Airplanes can land and take off by themselves, so why not do some trickle-down engineering towards cars? The more we build, the cheaper it will become” some might say. Well, the standards are extremely high and they must be. Would you want your plane to crash in the middle of an ocean? Components of airplanes are highly precise, over-engineered and require high labour cost parts which are checked and repaired to the most incredible detail possible by a highly trained workforce.

Compare these standards to the car industry right now. We can all agree there’s a huge quality difference. If it was about to change to higher standards, prices can only go one way: up, big time.

But let’s go a step further and imagine a perfect world where every single vehicle on the road is an AV. Vehicles will be able to interact between each other, share their knowledge and learn from experiences. All of this in order to eliminate the “human as a bad driver” factor. Beautiful idea. But how would you implement such a drastic change and how much time would it take to do so? And again, how much would it cost? Would they all be electric? If so, we would have to completely change our electricity supply system.

Another example: some drivers keep their old cars to go to work because they have no other choices. What would happen to them? There are so many economic, social, cultural and environmental questions that would need to be answered.

And yet, even if AVs could be cheap, extremely reliable, and have some form of intelligence and drive in all types of environments, what would happen when the so-called “trolley problem”? This is an extreme situation where the car’s software is forced during a potential crash scenario to choose between courses of actions that may cause harm to humans. As an example, a driverless car must choose between killing the two passengers or the three individuals walking their little puppy. Grim, but this scenario could potentially happen.

Should the car kill the two passengers or the three pedestrians? If the car was being driven by a person, they would definitely be blamed in this situation, whatever their choice is. But with an AV, would it be the manufacturer’s responsibility? Can car companies afford hugely expensive lawsuits against them? Would a manufacturer even begin to produce AVs, knowing all the challenges above? How can we deal with these issues, ethically and legally?

In our world full of imperfections and problems, how can AVs be implemented? Even in a theoretically utopian world where most conditions would favour AVs, quite a few questions are still to be answered. But let’s hope that science combined with engineering will, someday soon, solve these issues. Whatever happens, some of us will keep on driving our archaic old cars for quite a while.