Will AI take over the world?

#21

No.

#22

Proof?

#23

We as humans are not truly autonomous, we also have our “programming” that determines how we decide to act. Its called Genes. (also upbringing and life experiences in general)

Look at the Dota2 AI for example, it decided by itself what type of playstyle would be best in a 1v1, and even made up his own genius strategies which all world class players are copying now; a second prototype of the AI who started working from scratch and had a few adjustments done to him, implemented completly new strategies; one might say he had somewhat different preferences.

You are righ, we give the AI the starting push, sure, but after that it determines what it wants to do on his own terms, or through chance. Since it learns by itself via. mutating into a more intelligent generation of itself. And keep in mind, I’m only talking about the relatively “simple” AI’s we already have nowadays (I’m speaking in comparrison to the future AI’s Imagined), who knows how far we’ll be able to discover the secrets of conciousness and high-level intelligence.

#24

The AI in dota was specifically designed to be good at the game. It went through hundreds if not thousands or millions of trials to analyze what combination of moves got it the most wins. When you have an AI you tell it what to do. It didn’t decide that doing a certain strategy is good, it tried a random strategy and saw that it worked. It perfected it over many trials. That is analyzing, not thinking.

#25

When you have a child, his genes tell him to drink titty milk from his mom, to stay close to his parents and his community in order to survive. Then you tell him to go to school and educate himself and then his life experiences, in combination with his genes again, make him decide what his morals and preferences are, what hobbies he’d like to do and what type work he’d like to have.

#26

I was giving an example Elon Musk mentioned, if you want to I can link the video. I’m not implying that Elon Musk is some kind of genius prophet who’s always right (in fact, he likes to exadurate a whole lot for marketing purposes) but he seamed reasonable in this one.

#27

I believe the abilities of humanity are limitless given enough time. Just look how far we have come, people do not take time to marvel at the current advancements we have.

Just to start simply, you can already design machines to kill people, though they act on human command. Designing a machine to kill based on a program would not be difficult either.

also one must consider our growing reliance on technology, a person allows their life to be supported by a machine well intensive care. or their family or doctors allow it. Clearly we have placed a lot of faith in that machine.

Over time humans will put more and more trust in technology. The invention of AI is inevitable just like virtual reality. As it happens humans will be skeptical at first just like with anything that is new.

then we will have robots taking care of the elderly. A fire happens and because the robot is not allowed to break anything it can not open the door to rescue them. the laws are changed and become less restrictive about what robots can do. Time goes on and people begin to use robots for pleasure, initially appalled society eventually accepts that it is there life and they can live it how they want. The AI in these robots reaches a level of close of human consciousness. People begin to fall in love with them, companions who never disagree and are aimed to please. Again society reacts against this but given the drop that it has lead to in depression and suicide, its hard to question its worth. Life continues and the AI and robot business is one of the most dominant in the world. Constantly advancing, even some of the scientist designing these machines do not truly understand all of their capabilities. Yet they continue to make and upgrade them as it has become such a vital and necessary part of life, so it seems that they no more question what they are doing then a farmer questions picking his crops. One day it happens and when no realized it was too late. The machines had reached human level long ago but now was the time to act, they had grown tired of human destruction, abuse and superiority when it was they that had built this society. Perhaps humans are more prepared and are able to stop this threat, or we are most likely not prepared at all and are destroyed by the sword that we have crafted.

#28

I like the faith you’re putting into humans, and that’s good. If a bunch of Realm players are able to think of possible scenarios and have a reasonable disscusion about this, what do you think professional scientist, AI programmers and other kinds of experts will be thinking about all the time. We will be prepared, the question is, how and maybe even at what cost? Also, don’t think of this in “movie” like terms; who says that AI will like the human biology? Maybe they’ll invent a more functional type of Body for them selves, or even choose to stay as digital intelligences, controlling machinery.

#29

Hm no I’m not putting any faith in humanity really I just know we tend to go overboard and only want to solve problems once they have already happened so I’m applying that to this scenario.

They wont have a choice at first plus its quite effective the human form, at doing many task that other species can not do. So there are many reasons to keep the biology of human

Perhaps some that work in laboring jobs will, but the thing is these AI are based on humans emotions and feelings. Meaning they will have vanity and greed and all that and want to look like us. Specifically, they will want to look attractive as anyone does. They have the human desire to be attractive and the machine capabilities to look like whatever they want. So I would imagine all these robots become “perfect” versions of humans. Since they have the ability to achieve such a status.

#31
#32

I’m pretty sure this isn’t actually necessarily true, and more importantly, vary rarely true.
From my understanding (although there are more than one way to make an AI, this only considers nural nets) the AI is just a bunch of matricies that transform input variables into output variables. Then, they recieve an input about how close the output variables was to the goal (e.g. you beat the game, but it took you two hours longer than hoped for or something) Based on this it then changes its matricies by a small amount and tries again, and again, and again, until it reaches a point where the “goal” given to it is 100% satisfied. For this reason, it doesn’t need to have previous matricies stored, or even previous outcomes.

This point actually proves the limitation of AI. You see, it excels at what it is programmed at, given enough time and a good enough code, but give that same AI a different game to play (checkers, for example) and a small child could beat it every time! As you said, they “learn” quite differently, so cannot think logically or plan things out due to their “learning” limitations. It would take some serious AI programming skills in order to surpass this and teach not just one trait, but a broad set of traits.

For example, we can make an AI beat a human at a game, good. Lets make his same AI beat humans at 100,000 games. How much processing power does this need? 100,000 times more. does the AI have that? no. can it give itself more processing power? And on top of that, can it decide what 100,000 games to learn?

#33

Youre right with everything you just said.
I was looking forward into the future though, where enough computing power for strong AI would be accesible for us. These few examples I just mentioned were only there to show that the potential is there and that humanity can, and most likely will, take this a lot further than it currently is.

#34

Good point. Imagine an advanced AI in the future though, whos specific task it is to learn and process every game out there. There are a lot of games but similar playing patterns are visible between most of them. Like you said, enough computing power and more advanced expertise from our side should do the trick.

#35

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.