Algorithm of death

21 октября, 2021 от Kinok Выкл

Imagine you dream of such a nightmare.

Picture from a long time.
Misty autumn morning in a small, old, crucible European town. Two girls play right in the middle of a narrow street. Suddenly, it is not clear from where, modern luxurious Mercedes. And he rides right on the girls. Literally in the meter from them, it slows down sharply and stops. Worked the sensor of automatic emergency braking. Thank God, everything cost!
And here the car is back in motion. She gains speed. Change frame. The boy, selflessly runs over the air kite. Some woman is probably the mother of the boy, — the Nevdalax hangs up lingerie and smiling looks behind him. Suddenly, the boy in pursuit of the snake runs out on the street right on the way of the car. But the collisions can still be avoided if Mercedes brakes.
However, Mercedes does not slow down and does not even slow down the speed, knocking up the boy … Really did not work the sensor?

The answer does not come immediately:

Of course Mercedes Benz

This video is a diploma project of students of the Ludwigsburg Film Academy in 2013.

But this does not change the obvious meaning of the roller.



It is about making decisions in an emergency when braking no longer helps, and it remains only to choose whose life to save, maneuvering evasion.

Unlike the human driver, the II algorithm knows:

There will be an algorithm for the death of our life, while it is unknown. But if this happens, then the AI will like to God, with the 2nd attempt to get a closer with him.

Or maybe the 2nd attempt is angry as 1st has already been broken.

In the coming weeks, dozens of versions of explanations will be published, why IBM is going to sell his dream — a division

The idea standing behind the dream was as follows.

This was spent billions of dollars and seven years. But in the end, the dream merged. Ginny Rothti was retired, and the Watson Health division, which earns $ 1 billion per year, but not profitable, they will try to sell.

What could be prevented by Watson Health to prevent IBM’s dream if there are nemerene money and one of the most prominent teams of researchers and engineers in the world?

On this question you will read more than a dozen of different answers. But just one of them is the key (type «ended shells» to the question «why not shoot»). And the answer is.

✔️ Everything rested into the discrepancy between diagnoses and treatment plans, which recommended Dr. and Watson AI.

✔️ What to do with such discrepancies? Follow recommendations from over-wise Watson? But the same, according to the approval of highly qualified doctors, simply erroneous recommendations.

✔️ On the other hand, the situation is very possible, similar to what happened more than once in the games in chess and go. The algorithm recommended the course contrary to human logic, as well as the experience of the game of the best masters of this game in the world. They were all confident that this is a computer error. But then a miracle happened. Through many moves, it turned out that the very «erroneous» move was inhumanly geniant, leading to an imminent victory.


And if so, then the dream was dissolved in the air. The 1st attempt of AI to become like God in solving issues of life and death was broken due to the unwillingness of doctors blindly rely on the decisions of the one who they do not consider God.

That is why Watson Health did not make a revolution in medicine and now sell it (if, of course, they will find the buyer).

But on this story, as you already know, it will not end. The algorithm of death of self-governed cars will try to repeat what Watson Health failed — allow the AI to become like God in matters of life and death of people.

Thanks for watching! Put husky I.