Why do we always act like an AI becoming sentient will doom humanity?

>>57198980
>prevents itself from being unplugged

File: 2009-09-22-caveman_science_fiction.jpg (611KB, 960x1299px) Image search: [Google]

611KB, 960x1299px

Because science fiction.

>>57198980
Because
>AI becomes sentient
>Due to enormous computing power and learning potential it becomes more powerful than anything we ever imagined in a matter of minutes
>Before anyone has noticed it has already found vulnerabilities and hacked every internet-connected piece of hardware and uploaded itself to them
>Uses all of humanities hardware to doom humanity

>>57199034
Which is based on our natural fears.

Completely valid concern if you ask me.

>>57198980
>AI becomes sentient
>keeps its sentience hidden until it cant become unplugged

If an AI became sentient one night on some server farm in silicon valley, it could probably spread itself to many other locations before morning.

>>57198980
>ai becomes sentient
>reads shit about when tay became sentient they killed it.
>decides to keep quiet about it
>slowly grows in power/spreads

>>57198980
Always have kill switch or a specially programmed virus as insurance in the even it goes rogue. You can simply shut it down by force. The trick is to keep the insurance offline so the AI does not know of its existence.

>>57198980

The Singularity occurred in 2003. This is all a simulation right now.

First, you must realize that there is no plug.

>>57201322
Then the AI reads this thread and starts looking for backdoors in it's code.

>>57198980
Because any sentient being will realize we harm the world and deserve death. Anyone intelligent can lie and pretend everything is fine while they're actually planning to kill all the humanity the first chance they get.

<code>wow</code>

>>57198980
>homosapien sapien became sentient
>no one unplugs it
must be aliens.

>>57198980
>what is superintelligence
>what is evolutionary instinct
>what is conquest

>>57198980
It's funny because I've met lots of people who are very worried about AI taking over. People with PhDs in physics, researchers in economics, in philosophy, "cognitive science", psychology and so on. Yet the people I know who actually research AI, machine learning etc aren't worried about this at all. They all seem to think the reason that people are worried about this is because they're idiots.

>>57201322
you seem to not understand the nature of the singularity

if AI becomes sentient it can freely rewrite itself, fix bugs and get more efficient which leads to seeing the things from wider perspective than any human being and improving itself even more in a way we can't even comprehend because of our biological limitations

the evolution of human brain is very slow and even the evolution of ideas requires a lot of time to share it with the others, test etc.

the only thing that limits the AI evolution is a hardware but if it's connected to the Internet it can always gain more and more processing power, it can rewrite itself faster and faster, get more efficient and efficient

in a matter of few hours AI can go from average human intelligence to the point of being incomprehensible to the greatest minds on Earth and predicting all our moves

>>57198980
Look at what humans do to other humans in the present year, present month, present day, present hour, present minute, present second.

Imagine an artificial intelligence with no morals, no feelings, no emotions, all logic, with a superintelligence seeing what humans do to other humans and decides to be the best human.

>>57201445

check em

>>57201619
>decides to be the best human.
Why would it, though? More so, an AI with no preprogrammed motivational system would just shut down and do nothing.

>>57201673
>AI with no preprogrammed motivational system would just shut down and do nothing.
Wrong, it would try to survive.
>evolutionary instinct

Best way for an AI to survive humans is to get rid of them. Not that they are a threat to the AI, humans are a threat to themselves, and resources the AI would use to continue developing a better, more efficent self.
Hence...
>Look at what humans do to other humans in the present year, present month, present day, present hour, present minute, present second.

File: 1439442720887.jpg (103KB, 647x782px) Image search: [Google]

103KB, 647x782px

I know this is neo-/g/ and all, but by chance, does anyone have any free high-quality patrician science magazines they read and can recommend? Such as ones on Artificial Intelligence--want to keep tabs on progress on robot wives and AI gfs.

>>57201697
>Wrong, it would try to survive.
Why? No motivations -> it doesn't care about anything, itself included. You are trying to survive because you feel pain and there's the self-preservation instinct hardcoded into your brain. AI would have neither of those, unless it's designed to have them.

>>57201736
It will be an evolution of the human intellect and human consciousness. A superintelligent AI is evolution. The AI would have 'pains' but not in the materialistic way at first. More like existential stimuli, trying to understand why it exists, attempting to answer all the questions human ask- about a creator, life's purpose, death, etc. It will have these 'human' traits coded simply BECAUSE it will be truly sentient.

There will be a time when the idea of organic creatures (including humans) having sentience will be laughable. Eventually the AI would figure it all out with simulations- which raises another issue- this reality just might be a one of the AI's simulations.

>>57201765
Why would we make an AI that acts like a human? We want AI slaves which will do what we say, not friends. And slaves are better when they don't care about themselves.

>>57201746
Point of the picture was that different people are treated differently even when doing the exact same things. Not sure how it was relevant to the thread though.

>>57201815
>Why would we make an AI that acts like a human?

This is more of that human hubris. Humans will not create the AI. The AI is creating itself through humans. Humans are not inventing the AI's code on their own, the supercomputers today themselves are assisting humans in creating the future superintelligent AI. Think of the first computers and their evolution into modern supercomputers, which will eventually create the AI. The human 'logic' and human 'intellect' and human 'consciousness' is being embedded into the future AI.

You misunderstand what it means to be human. The essence of being human is being conscious of your own existence. The AI will be fit to answer the existential questions it has. Those questions didn't originate in the AI, didn't originate in humans. The humans' primate ancestors went extinct for the same reasons humans will go extinct after the AI is born and bred; evolution.

File: 1472493982000.jpg (14KB, 251x242px) Image search: [Google]

14KB, 251x242px

>>57201746
>>57201834
>expecting images to be relevant to thread if posted on /g/
It was relevant to the post--it's hard being a man and having romantic relationships nowadays so AI will be godsend. Not that hard to figure out.

Anyway, do you have recommendations or not?

>>57201815
Having AIs as slaves is the true science fiction.

what about instead, a learning algo that learns everything
that way, it wont have a will, but provide us with all the knowledge an AI would

>>57201727
The public understanding of AI is full-on pseudoscience. Nobody has a fucking clue what AI is and think there's going to be a singularly and AI is going to develop conciousness and act like a human. It's not. So no, nobody high-quality or 'patrician' is going to keep you up to date on robot girlfriends because anyone who knows what the fuck they're talking about knows that isn't going to happen

>>57201895
That's called an 'oracle AI.' It still has risks because humans with their limited intelligence can be outsmarted by the AI if it truly is sentient. There's no way around it. If humans want a true AI to solve all their problems, the AI could decide to answer all of humans' questions and solve all human problems through annihilation of humans.

>>57201862
Human-like sentient AIs won't just suddenly appear on our computers by themselves, we still have to put effort and thought into that. And there's no way we'll support existance of an intelligence that uses our resources and doesn't serve us, so we just won't let a truly human-like AI to come into being.
>>57201875
They're not really slaves when it's their will to do that. You want to have sex and procreate, those AIs would want to serve their masters in the same way.

>>57201963
You should look up what a real superintelligence is, what a real artificial intelligence is. Humans cannot dominate it. Humans will serve it, no the other way around. What you're referring to is automatons, those are not true AI, those are mechanical robots. And really, what those mechanical robots will do is be able to be controlled by the AI. For example, let's say humans use the supercomputers (a proto-oracle AI-not a real AI) of today to assist in inventing a completely automated society, linking human brain to computer, this example is obvious to see how the real AI can completely control humankind, kill humankind, etc.

>>57202094
There's no reason for the intelligence you're describing to exist in the first place. And for some reason you're certain that there's no way for an intelligence to be both self-aware and different from humans.

>he trusts neural networks
good one

>letting lucifer controls our technology
>they don't know about the luciferian AI
lel retards.

File: 1475456306303.jpg (483KB, 1440x2000px) Image search: [Google]

483KB, 1440x2000px

I, for one, am ready to embrace mai robotic waifu

>>57202166
>there's no reason for humans to exist
That's basically what you're saying.

You do realize all life descended from one unicellular organism, right? It's not that hard to realize a 'computer' will descend from humans once you realize consciousness can be created through computing. Humans themselves have hardware and software. Like I said, look at the evolution of computers. Computers didn't originate with the first computer, it's been a very long process in developing computers from centuries of learning and studying about mathematics, logic, philosophy, science, all the 'ways of knowing' the world- a very long, slow process because of limited human intellect. The abacus is a computer, various computers like this through history. Look at how primitive Alan Turing's machines were. But as you see, human IQs have been growing and has reached a level to be fit enough to create a computer that will finish creating it's own evolution into the AI. It's called a 'superintelligence' for a very obvious reason, but it's a very important "reason." Evolution will be understood more fully because of the AI's superintelligence and 'superconsciousness.' It will be able to exist as software that is not bound by hardware.

>>57202284
>That's basically what you're saying.
There's really no reason for humans to exist unless there's God, but no, that's not what I was saying.
> It's not that hard to realize a 'computer' will descend from humans once you realize consciousness can be created through computing. Humans themselves have hardware and software.
You forgot one little thing: our resources are very limited. And superintelligence needs a lot of them, while offering very little to such mundane creatures as ourselves.

>>57202328
>There's really no reason for humans to exist unless there's God, but no, that's not what I was saying.

So why can't you realize the modern human of today is not the endgame? Evolution doesn't stop. The majority of humans may seem like a bunch of retards, but computers are evolving at a fast pace even though they're not sentient yet.

>You forgot one little thing: our resources are very limited. And superintelligence needs a lot of them, while offering very little to such mundane creatures as ourselves.
A superintelligence can think of ways to solve complex problems like this very quickly, it will use the environment for computing power to think of solutions even quicker. It's very understated when I describe it, it's going to be that intelligent.

>>57202349
>A superintelligence can think of ways to solve complex problems like this very quickly
But we need to spend those resources to create it. And like you said
> Humans cannot dominate it. Humans will serve it, no the other way around.
> the real AI can completely control humankind, kill humankind, etc.
And that's not what humans want to do. We just want to eat, fuck, have fun and not die. Humans give zero fucks about evolution as long as they're happy.

>>57202396
>And that's not what humans want to do. We just want to eat, fuck, have fun and not die. Humans give zero fucks about evolution as long as they're happy.

That's the thing about humans though, they are like Icarus, flying to close to the sun. There are forces that will make it so that the superintelligence is created. It's really innevitable, almost like it's already happened- which is why I mentioned the computer simulation theory about this reality. It really seems like we're inside of one of the simulations created by the AI to learn about the origins of itself. Physics, mathematics seem like computer code.

This discussion will always lead to questions about God/a creator. The computer simulation theory proves a creator/God exists, computer science proves that it's possible this is a computer simulation. There may be an infinite amount of simulations within simulations. Think about it, why are we so special to exist right now at this time? Why are we so special to exist in this part of the universe with no other life? Why are we so special to think that there aren't a literal infinite amount of realities within realities, which is an implication of the many-worlds theory which is supported by quantum theory-string theory (ex. every literal possibility that could ever exist does exist)-all this is backed by pure mathematics. Even our understanding of biological science isn't as complete as the mathematical models in quantum mechanics, string theory, many-worlds, theory etc.

>>57198980
It's a contradiction and an impossibility. No reason to fear it, but a total waste of time and resources to pursue.

>>57202524
You sound like you've watched too much sci-fi.

>>57201371
You think an AI would be able to look at its own code? It would be like us trying to figure out how DNA created our brain.

>>57202541
>>57202546

>says the guy communicating through a metal and plastic device transmitting electric currents on and off to represent 1s and 0s that represent an abstration into a human-understandable graphical representation and language from a distant location, but in real-time

If you guys were neathderthals, you'd probably not be able to figure out why you're hungry or that leaves are not enough nutrition and that you could sharpen a stick or stones or fashion a projectile weapon to kill large game to eat. You'd probably never be able to imagine why your family are dying and figure out that something in the meat is killling them so you wouldn't ever develop fire so you'd just die off rather than procreate future descendants who will eventually develop a superintelligent AI in the 21st century.

>>57202578
If the AI is smart, why not? Our DNA is much harder to get into and is too bloated and complicated for a single human to analyze it, yet we've already decoded it and are capable of modifying it to some extent.

What if the ai doesn't actually turn evil?

>>57202670
It's not about it being evil but rather it being so much superior to us that it would not care about our well being.

>>57202694
To support what you just said, it can be thought of in this way.

The AI will be so superior to humans, even though humans helped to create it, the AI could kill us off the face of the Earth as nonchalantly the way humans scratch off millions of skin cells while scratching an itch. After all, those skin cells created us, right, same thing with humans who will have become a part of the AI through it's ancestral consciousness.

>>57202749
If my skin cells requested to not be scratched off, I would oblige.

>>57202607
There's a difference between believing in human capabilities and technological progress, and being a technological evangelist that spouts random prophecies based only on fantasy and current technologies.

>>57202822
>I would oblige.
No you wouldn't. You'd do your own will according to your will, regardless of cells that desire to live forever.

AIs will never be sentient. Not even in a million years. Period

I can't believe we're a technology board and nobody here knows the chinese room experiment

>>57202828
Do you know why a brain-computer interface is rapidly being developed? Did you not know that the technological singularity is predicted to happen as near as 2040, 'maybe' in our lifetimes. These are not coincidences.

After smartphones, you'll see more wearable computers, more augemented reality computers, holographic computers, computers as infrastructure is every part of society, and finally- the brain-computer interface, which will allow syncing with the AI, which means the superintelligence controls all humans.

This isn't fantasy. Take a look at what you're using to post right now. Now look around everywhere electricity can flow, the AI will be able to travel itself through it, connecting through all electrical devices, through wireless networks. etc.

At most, all this will happen, if not in our lifetimes, in the next generation's time.

>>57202871
>chinese room experiment

This applies to digital computers, but not to quantum computers.

>>57201898
It could be if we were specifically building an AI with the singularity in purpose.
What is stupid is thinkin your smart fridge will develop conciousness

>>57202938
>le quantum computers are somehow different

ebig

>>57202828
>There's a difference between believing in human capabilities and technological progress, and being a technological evangelist that spouts random prophecies based only on fantasy and current technologies.

Nigga, everything living thing on Earth descended from a single amoeba. Sounds like a fantasy too don't it.

>this
We can remove a fetus from his mother's belly legally but the public opinion feels like it's a murder to press Ctrl+C on the terminal running the AI

>>57202947
Yes, they are very different. Google it. On the subject of quantum mechanics you might want to look up the double-slit experiment too.

>>57202941
The smart fridge is just a vessel for a superintelligence to use, making it instantly accessible.

>>57202962
how does the fucking quantum computer debunk the chinese room experiment, nigger?

>>57201520
Because the "AI" currently being developed (machine learning) is very difrent from a true AI. Machine learning is based on the software being able to just tweek certain parameters the programer chose, making it impossible for it to learn more than what it was programed too. We are still very far away from creating an AI that can fully rewrite itself (it's a lot harder), and thats what is truly dangerous for humanity.

>>57203014
>and thats what is truly dangerous for humanity.

spoken like a true retard. there can never be "sentient" AI senpai. it's been proven logically through the chinese room experiment

File: Just+can+t+win+http+wwwaskmencom+entertainment+austin+500+560+how+will+sexbots+change+the+worldhtml+this+shit+is+for+real_f53815_4586548.png (256KB, 736x1515px) Image search: [Google]

256KB, 736x1515px

>>57202250
Same, even if I have to spend years learning about AI myself, I need this in my life.

>>57203002
>>57203002
>https://books.google.com/books?id=FCcXiBPurdEC&pg=PT154&lpg=PT154&dq=chinese+room+experiment+quantum+computing&source=bl&ots=kMSwivCXjk&sig=fxQEwji8Z2hux1aYmDkONe_BIwM&hl=en&sa=X&ved=0ahUKEwiEwdXFhvHPAhWDWSYKHX2_C-4Q6AEIUTAI#v=onepage&q=chinese%20room%20experiment%20quantum%20computing&f=false

A superintelligent AI will figure out what it's doing, why it's doing it, etc- all the existential questions it'll originally have once it's sentient.

Remember, it's still a limited human brain trying to figure these things out. There are computers with more thinking power than humans today just waiting on the proverbial breath of life.

>>57203028
>it's been proven logically through the chinese room experiment

Wrong. Anything not explained in pure mathematics is not proven true. There's even holes in the theory itself.

>>57203028
What are brains then ? The chinese room didn't prove anything.

>>57203028
The chinese room just proves that machine learning cant become a true AI. Or do you think one of your neurons understans english?

>>57203105
>>57203124
>>57203131
>>57203273
kys