Do you think that AI will ever feel emotions?

But neurons are merely a communuication network. The information is bio-chemical.

It's still chemical, produced by glands and transported via the cytoskeleton. Can this very sophisticated neural system be replicated in AI?

 

Log in or Sign up to hide all adverts.

That is the question, no? Can the mammalian internal neural communication network be artificially copied?

 

Log in or Sign up to hide all adverts.

It occurred to me that our "education" of an AI is actually very rudimentary. Human children have to attend school for some 15-18 years before they are set free on society. This is not just learning by rote, but also by social interaction.
I wonder if an AI needs to learn social skills?

 

Log in or Sign up to hide all adverts.

It will be able to give a convincing impression of doing so. After all, autistic people have to learn - intellectually comprehend - the feelings other people express naturally. AI has a capacity for learning, the limits of which we have no way to predict as yet. It can master very complex operations. Reading human (and for that matter, canine, equine, feline...) body language, facial expression and voice inflection would be well within its capability.

I don't think so. Actually, that sounds like telepathy. Between humans, even humans of the same culture, feelings have to be expressed in the ordinary forms of physical communication: we make appropriate sounds and gestures.

No, and I don't think there is any point in trying. AI is a different life-form that works on different mechanical principles. It didn't need billions of years of trial and error - and catastrophic failures - to build an information-processing organ, and its more efficient: not lumbered with all the vestigial obsolescence we carry from generation to generation.

But a kindergartener needs an hour to learn all the words to Twinkle, Twinkle, Little Star, while AI learns the Iliad in one second; in ten, it will have translated the text into seven modern languages, with commentary.
Also, like humans, it continues to learn from experience, including social interaction. Only faster. Plus, it won't be inhibited by human shame or fear of censure in seeking the information it requires for a new operation.

Yes, absolutely!

This is interesting https://towardsdatascience.com/artificial-intelligence-and-autism-743e67ce0ee4

 

Seems like raising the bar. We know ai is still in a rudimentary state, it's not general intelligence, but tightly focused. But I see no theoretical barrier to general intelligence. Human brains can do it. When we come to understand how that works, we will know better how to create a generalized ai. Or maybe we will just scan a human brain and imitate its functions. Note that autistic people are human, intelligent, and have emotions, even with their empathy and social issues. Learning social skills by rote, as many autistic people do, is not necessarily a bad thing.

 

One of my first thoughts here was about autistic children. This thread made me look into that. I was quite astonished to learn that it's AI teaching the autistic people how to respond to feeling communication. It's way ahead of when I last checked!

 

Do you think that AI will ever feel emotions?

... do you think humans are normal if they are not mentally and physically affected by emotions ?
would you trust humans judgement and intentions if they were not effected by emotions ?

would you work for a boss who had no emotions ?

 

clapping monkey audiences annoy me
but thats american western culture

 

Not in my example. I am talking about replacing neurons in your brain with electrical prostheses. No biochemical signals, just electrical. (Which works just fine to stimulate nearby nerves - and in fact this is currently done for hearing and visual prostheses, and for treatment of things like Parkinson's.) I contend that if the prosthetic network works exactly the same as the original network (same inputs give same outputs) you will notice no difference.

Anything can be simulated in a computer. If you simulate your brain and body with 100% fidelity, would there be any way for you to tell you are not "you?" (Note that we are decades away from having the computational power to do this to even a fair degree of fidelity.)

 

Duplicate

 

AI twins ?

Employee "im not feeling very well i have a bad feeling about this business idea"
AI Boss "feelings are subject to perk tax & that is not in your salary package you are fired!"

: Employees AI sex doll emotional companion union rep replys, "im getting a bad vibe about this"

... AI world ... no unemployment benefits... only civil war

 

what sides?

 

Is this what you are referring to?

I have a feeling that electricity is not sufficient by itself. IMO, the phenomenon of consciousness requires electro-chemical communication at quantum scale.

But would it be easy to test on a single celled organism. They are relatively simple and have very limited capabilities.

Any comments?

 

Why?
That's been true within our experience - which is limited to evolution on one planet out of all the unimaginably big universe. We are only aware of organic consciousness, yes. But why do you think there can't be a non-organic kind?
(I'm inclined - instinctively - to agree, but have no scientific/rational basis for that belief. Do you?)

 

Holistically they don't . All cells communicate with all other cells .

As I have before life intelligence is much different from electronic intelligence .

No electronics can form without Life giving it form .

 

I agree in principle, but also believe that the most efficient evolved complex patterns have been shaped and honed by the evolutionary process. If the human mind is an example of an efficient evolved organ and it uses electro-chemicals as its most efficient or most functional combination for complex information sharing, then that should teach us something about the combinatory richness and functionality of electro-chemical information transmission, rather than just electrical or chemical assets alone.

 

In other words: anthropoconceit.

The one thing evolution isn't, or ever has been, is efficient. It's all been trial-and-error, ad hoc, hit-and-miss, obsolete semi-functional mechanism patched with repurposed components, system piled on system. The human brain, in particular, is running on spaghetti code - and malfunctioning with terrifying frequency.

Then why is it able to learn so slowly and forget so quickly, compared to a computer?

That just mean the materials available to nature 3.5 billion years ago worked on those principles. If that design had been intelligent (or a design), it wouldn't have taken so long. Now, something exists that's capable of producing intelligently - or at least purposefully - designed information systems, and also has available materials that didn't exist the first time around.

 

No, all organism that have managed to survive are equally capable of surviving. Their existence is the proof. From a survival perspective, the insect is the most successful organism on earth, aside from the bacteria.

Well, that's debatable in the long run. Natural selection is causal to ever increasing sophistication (adaptation) in organisms, but there is a price to pay. The same holds true for AI.

To make room. But you are wrong in the ability of the human mind to learn slower as compared to a computer.
Speech recognition is one area in which the human brain is extremely flexible, whereas a computer is considerably more rigid within a set of fixed parameters. You can clearly detect a delay in an AI's response to a question which humans can respond to immediately. The AI can gather data at incredible speed, but using that data is a whole other phenomenon.

Check this little but excellent presentation by Anil Seth and his demonstration of human cognitive powers. It's really remarkable how quickly the brain can learn, but as it is guessing at what is being observed, it can be fooled as well. For examples of brain flexibility and weakness, start lecture @ 12.45

Really? The insect had conquered flight some 350 million years before man crashed his first attempt at flight. And we had the birds as examples. Would you say migrating birds are not perfectly built for long distance flight and navigation? How many single persons have crossed entire oceans or continents. Millions of birds do it every year.

That's a perfectly relative statement. Yes, it is capable of constructing a baby machine when materials are available.
OTOH,
Biological organisms need no materials to construct a baby machine at all. They grow them! If you believe that AI can outperform nature in growth, evolution, natural selection over a range of states from the very subtle to gross expression in reality that is limited thinking.

Sure, nature makes imperfect choices but imperfection tends to get selected "out" eventually. How many planes did the Wright brothers crash? We even had birds as examples and proof that flight can be copied.

Natura Artis Magistra. (Nature is the teacher of the Arts and Sciences)

 

You don't even know why, so we can't say it's necessary. In the end, the output is information, I doubt it matters what medium that takes. One difference is it's analog not digital, what if we made an analog computer?

 

In the end the input is information, the output is guesswork by the brain. And if the brain has quantum functions it matters a lot what medium it requires to process information. Check out Dr. Stuart Hameroff, he does know why certain things are necessary for thought in biological organisms.
When you have an operation, he's the guy that turns you into an object and back again into a human.