When it comes to artificial intelligence, or AI, we usually think of robots that look like people or manpower replaced in performing complex tasks. Development is expected to be limitless, but how much do we actually know about this field?
First, it is not a settled science. It is unpredictable and often accused of giving less than expected. Apocalyptic movies often show that human race could be overtaken by superintelligence, getting more powerful over time. The phenomenon, called existential risk from artificial general intelligence, has been widely debated, but without conclusion.
“A curious aspect of the theory of evolution is that everybody thinks he understands it,” biologist Jacques Monod said.
The risk means that intense development of AI will cause human extinction. In 1965 I. J. Good introduced the “intelligence explosion” concept, later known as “technological singularity”. It assumes that machines would become smart enough to continue their progress by themselves. Machines would create even smarter machines, leaving the intelligence of the human race significantly behind their own. Many famous physicists, from Stephen Hawking to Roman Yampolskiy, and entrepreneurs such as Bill Gates, expressed their concern about the way that superintelligence expands.
Robot rebellion where AI spontaneously develops some habits to break down the world order exists only in sci-fi movies. The real danger comes from human dependence on scarce resources, which robots could use in a more productive way if they develop the ability of making high-quality decisions.
The second threat comes from the inability to have moral values. The lack of ethical standards could make machines do whatever it takes to accomplish a goal, even at the cost of human life.
Artificial intelligence, a system that can overcome human capabilities in every relevant area and easily defeat people whenever its goals conflict with human ones, will lead to extinction of the human race.
It is hard to estimate the rate of improvement, but some scientists predict that it could be high. While some claim that this is a very possible scenario, others think that the thesis is crypto-religious, where superintelligence replaces God. There are those who think that human race will already destroy itself before the progress of artificial intelligence accelerates . However, both sides agree that regulation in this research field would be pointless.