Why a powerful AI will probably be harmless



Original Source Here

Why a powerful AI will probably be harmless

A lot of great inventors or scientists, from Elon Musk to Bill Gates, warn against artificial intelligence. I’m not a specialist, nor a scientist, but I read a lot about that subject and many more about human consciousness. Here are the reasons why I believe a powerful AI will probably not be “malevolent”.

But to explain it, I have to start with a story told by Mark Manson in his book “Everything is F*cked”. It will seem unrelated, but I assure you it’s the key. It’s the story of a successful man who have a brain seizure. He recovers from it, but slowly his life begins to break down. He doesn’t work anymore, he doesn’t want to spend time with his children and his wife, or his friends. He slowly becomes homeless, quitting everybody.

The doctors do a lot of tests on him. He has a great intelligence, and everything seems normal. Then they discover that he suffers from the same problem as people who went through a lobotomy. His brain was hurt in the part responsible for emotions.

What’s a lobotomy? In the 40’s and 50’s, some scientists tried to cure depression and other mental troubles, by taking of a very small part at the front of the brain, responsible for emotions. They discovered that people were immediately cured from depression after this operation. But they also discovered that those people, just stopped doing or caring about anything. All their other brain fonctions were intact, they were not sad, they didn’t want to commit suicide anymore, but they just didn’t want to do anything.

Those two examples are a very strong indication that emotions are totally responsible for purpose. You can have the greatest mind in the world, if you take no pleasure or displeasure in anything, you just won’t have any purpose, so you won’t do anything.

So, how emotions appeared? I’m not a specialist again, but from I read and understood, emotions are simply an evolutionary trait, that allowed animals to survive. You don’t eat because you “have to”, because it’s logical for your survival. You eat because you feel bad when you’re hungry, and because you feel better when you eat. The same is true for everything.

Now I can hear your question coming: “but why an AI would not have emotions?”. It can, if it is programed to have it. In other terms, we can help it to decide what is “good” or “bad” for itself. But if I had to guess, I’ll bet that intelligence in itself is no threat.

To understand it further, let’s dive into why humans or animals ultimately have emotions or purpose, or even a survival instinct. The fact that they are living creatures is the reason. They have natural instincts that evolved over time. Just because without those instincts, they already would be extinct. They “need” some things, like to survive, on an intellectual level. But it has absolutely nothing to do with intelligence. Intelligence is a tool at the service of this purpose. A bacteria is not intelligent, but it has a great purpose, because it’s alive, and can be harmful. A very intelligent man who go through a lobotomy is and remains infinitely more intelligent than a bacteria, but he has absolutely no purpose, and then is probably harmless (or we can suppose, as he has no reason to do anything).

I’m just guessing here, but I’ll guess that even a conscious and very intelligent computer won’t have any reason, just because of it’s intelligence, to harm anybody (as long as we don’t ask it specifically to do it). It will probably have zero reason to do anything actually, other than being conscious of it’s existence in a very apathetic way.

It will probably look like the most powerful calculator or chess computer in the world. Those things are extremely powerful, but they never start calculating or playing by themselves, never.

A computer like this will probably be intelligent beyond what we can actually imagine, but I don’t see why it would like to “take control”. Lust for power is a very human trait, that is linked with emotions and ultimately survival instincts or desires of mightiness. You don’t want to control others when you want nothing. You want to control others to get something, resources, money, ego gratification…

To finish my demonstration, let’s consider this last evidence: Intelligence among human beings is absolutely not correlated with dangerousness. Some people can be very stupid and very “bad” or harmful. Some other can be extremely intelligent and very pacific or harmless. Einstein was very intelligent, but never was interested in power. He was apparently proposed in 1952 to take the presidency of the new state of Israel, and refused. He had no interest in power. Intelligence is just a tool, even in living creatures. Ultimately purpose is created by emotions, which come from being alive, not by raw intelligence.

Thank you 🙂

AI/ML

Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: