If AI “explodes in intelligence”, are ordinary humans only ants or pets?


liu, tempo Date: 2021-08-02 09:23:29 From:ozmca.com
Views:102 Reply:0

Imagine that when you knock down the dominoes, the first card falls, and then each card reacts quickly until the last card. In fact, this chain reaction is not limited to the physical world, but widely exists in various scientific fields.

 

Today, scientists are discussing a new field that will produce chain reactions-AI. Many people have proposed that AI’s intelligence is like dominoes. Once a switch is turned on, it will constantly replenish itself, forming an AI “intelligence explosion”, and may eventually surpass human intelligence.

 

Recently, the world-renowned artificial intelligence expert and former University of California professor Lance B. Dr. Eliot published an article in Forbes magazine, discussing the issue of AI “intelligence explosion” in detail.

 

What you are talking about – “AI” is not “real AI”

 

Some scientists believe that one day, AI may have an “intelligence explosion” and claim that this may happen between 2050 and 2100. But no one knows the specific answer.

 

In fact, the existing AI systems cannot be compared with human intelligence. The current computer-based AI can do some very tricky and impressive things, but at best it can only be called “narrow AI”-it can only work in a limited range or field, and lacks interaction with people. Common sense reasoning ability in dealing with.

 

real AI

 

The real AI, usually called AGI (Artificial General Intelligence), can pass the Turing Test without any suspense.

 

But currently, there is no AGI in the world.

 

Some people believe that only humans can display the highest level of intelligence. Therefore, no matter how much AI intelligence evolves and develops, it is impossible to reach a level higher than that of humans. Therefore, after an intellectual explosion, AI will reach a level that is completely equivalent to human intelligence, and then stagnate.

 

But on the contrary, there is another view that the explosive growth of AI intelligence will surpass human intelligence and stay at the level of so-called super intelligence.

 

How high is the level of super intelligence? We have no idea about this for the time being.

 

But some people firmly believe that although superintelligence will exceed the level of human intelligence, it will also have a limit or ultimate advantage in itself. At the same time, the explosion of AI intelligence may be an endless chain reaction-AI may continue to develop, triggering super AI, and then super super AI… In short, becoming smarter and smarter forever.

 

AI intelligence explodes, is it pros or cons for humans?

 

One question we have to pay attention to is: Super AI can think, do, see or say what we ordinary people cannot do?

 

The answer given by Professor Eliot is that superintelligence can probably solve cancer, world hunger, explain how the universe started, and in essence can answer all the thorny open questions that humans have only perfunctory intelligence that has not yet been solved.

 

The sad thing is that ordinary human intelligence may not even be enough to recognize what superintelligence can think of. Therefore, by then, humans may be like little ants in the eyes of AI. Super intelligent AI will either tolerate our existence or help we survive.

 

It may try to push us to the intelligence ladder, or decide to make us “pets”, it may teach us new wisdom, or it may think that we are not worthy of its troubles and give us up.

But on the other hand, does this mean that we don’t need any other new inventions at all, because in some way having superintelligence and AI can meet all our needs?

 

Imagine that sometimes our intelligence is improved when we are with someone who is truly smart, because the interaction with the other person may make us addicted to, absorb, or amplify our intelligence in some way. Therefore, if we achieve AI superintelligence, it may lead to an increase in our own intelligence as a species, which seems reasonable.

Leave a comment

You must Register or Login to post a comment.