My article in BrandsIT dated June 27th, 2018.
Developers involved in creating algorithms used in autonomous cars should be prepared for the fact that their work may soon become the subject of heated ethical debate. The growing popularity of autonomous vehicles is already occupied by the heads of lawyers and legislators. This is because the issues of good and evil are complicated, related to road safety.
Car industry is closely watching the achievements of Tesla and Google. The latter regularly provides reports, which show that its autonomous cars are traveling more and more kilometers without a fault and without creating dangerous situations for traffic. Noticeable is the growing optimism of the entire automotive industry when it comes to the future of this type of car. Their mass production will start or run soon: Mercedes, Ford, Citroen, Volvo.
Almost every major manufacturer can build an autonomous vehicle today, test it and prepare it for traffic. But that’s just the beginning. For some time, the demand for cars and revenues of corporations due to their sale will be limited due to complicated legal and … ethical issues.
On the way, we see a car driven by a man. A wife is sitting next to him, two children in the back. Suddenly, a child runs out in front of the vehicle, chasing after the ball. The distance between the child and the car is minimal. What does the driver do reflexively? Either he brakes suddenly or violently turns … and lands on a tree, wall, in a ditch, and in the least optimistic scenario he enters other pedestrians. And what happens when there is an autonomous vehicle in the same situation? Imagine that his driver, satisfied that he does not have to worry about anything, took a minute nap. The car must choose what to do when a child pops up on the road. What does the car do? We do not know. It depends on how it has been programmed.
In short, you can theoretically choose three ways to – let’s call it – ethical programming algorithms that can largely determine the behavior of the car.
The first way assumes that in the case of an accident and the threat to life, the good of all the participants of the accident (ie the driver, passengers and the child on the road) is jointly and severally important.
In the second approach, the most important is the life of pedestrians and road users.
In the third approach, the priority is to protect the life of the driver and passengers.
Everything depends on what algorithm will be decisive in this case and which one will choose the given company.