Experience emotion?
Seems more of a very bad translation to cover plagiarism than a modern AI product.

https://archive.ph/jKJHn
Brandon Hunter useless at 42

Former NBA participant Brandon Hunter, who beforehand performed for the Boston Celtics and Orlando Magic, has handed away on the age of 42, as introduced by Ohio males’s basketball coach Jeff Boals on Tuesday.
Hunter, initially a extremely regarded high school basketball participant in Cincinnati, achieved vital success as a ahead for the Bobcats.

https://futurism.com/msn-ai-brandon-hunter-useless

 

Unpacking this is a bit tough, but lets have a go...



"I mean faked by way of a program..."

How do you know that "real" is? From later in your post you seem to tie emotions to particular physical ways of producing them. If you are saying "real emotions can only come from the brain structures involved in manifesting them in humans" you are begging the question with a circular definition and that's not very useful.

Start by defining "emotion" in the "real" sense. What is a "real emotion"?

You go on to say "a reply that fits what another human would expect" but can you actually distinguish this from how a human acts as a result of emotional provocation? But this simple sentence is a dense web of assumptions so to go further...

"...a program that knows..."

Some irony there. If the program knows isn't that intelligence? If you can even talk about a program "knowing" isn't that an indication of the problem here? How do you distinguish between "knowing" and "programmed response"? Why do you think that Turing, when devising his test, only measured the phenomenological aspect and didn't concern himself with a code review or the methodologies of the programmer?

You've run into a serious epistemological dilemma here—so many smuggled assumptions. You really need to back up and provide some definitions.

"It doesn't actually *feel* anything."

What does "feel" mean?

'Even if it said it "feels bad about that" it has to be faked, or programmed to respond in that way.'

"Faked" or "programmed to respond in that way"? How is "faking" distinct? Won't a "real emotion" be "programmed" since that's how computers are configured to do things?

I am assuming you mean something like a lookup table of "appropriate" responses. "When A is detected respond with B", sort of thing. The trouble is, the current SoTA for AI does not involved procedural algorithms of that sort. It involves complex neural networks that so rapidly grow in complexity the "programmers" have no idea how the problems are being solved. They can't reverse engineering the networks, and they can't use the results as a basis for a new version they optimize by hand.

The core of this technology is Machine Learning. This is based on various ways of turning raw data into structures that manipulate that data. In many ways, it is as mysterious as "real emotion", and it is much faster than anything biological in evolving. So...

"It's doubtful that it has the deep emotional capabilities of a human because according to evolutionary theory that stuff came after thousands or even millions of years."

Notwithstanding the anticipation of a definition of "deep emotional capabilities", your concern about time may not be a valid one.

And, one more point—the problem of "experience" is a fraught one. The necessity that an AI "experience" the world as we do in order to be "truly intelligent" is, in the absence of some theory of experience itself, purely arbitrary. You can know what it is "like" to be you, but you can't know if that means knowing what is it like comes from a separate thing from being you.

In fact, it appears that it really can't be something separate from everything that makes you act as you do. The unmeasurable and unobservable (by others) nature of experience makes it opaque—hence the phenomenological requirement in dealing with it. We can only see and measure, and therefore test, what is exposed to the world.

This particular question requires the most rigorous philosophical approach to even decide what the question itself is. The vast extent of the assumptions you are making in your attempt to "answer" a question that, itself, isn't even clearer a valid question, means that you have no chance of exploring the underlying "truth".

Start with definitions. Question your assumptions. You are not very likely to come up with "answers" but if you manage to come up with questions it will be a big win on the road to understanding the problem you are—at this point—glibly waving away as obviously the way you assume things are.

(N.B.: None of the foregoing is intended to be personal, belittling, or critical—I am responding from the point of view of the philosopher examining an answer to a philosophical question, and, while many people focused on "hard" sciences don't recognize the rigor of actual philosophical investigation, it not only exists, it's something that requires a lot of work to both understand and do.)

Hi there Ya'akov,

Some very good counter points, I thank you for that.

The problem you outlined is certainly valid. The problem for me is I don't think I can take the time to write out a 100 page dissertation on the abilities of an AI program (or whatever you like to call it) versus the human capabilities. It's just too complex as I think you understand completely.
All I can do to start with is outline my own 'feelings' on this. I do have some background in machine learning, however my knowledge on the more modern ones may be lacking.

What I can do though is summarize one of the key points, but unfortunately nobody has the answer to this yet. That still leaves us in the dark, but it may lead to a turning point in the decision of whether it is really feeling emotion or just saying it does (if it does say that, that is, which it does not seem to do, and if we question its own response then that is only because we can question anybody's response and if they are really lying we can't tell that either, so little point in that also).

The first bottom line as I see it so far and as I was saying before, could be whether or not the program or whatever you like to call it can work via some sort of quantum effects, if in fact it is ever proved that the human brain does rely on quantum effects. That could be part of what makes it so complex. I can't say for sure of course so we just have to wait and see.

The second bottom line is, and maybe this is even more significant, could we be seeing someday the meld between biological and electromechanical 'technology'? It may be possible that the latter becomes equal to the former and thus we will see true sentient beings emerge.

The other thing to remember is that none of us has the perfect answer yet, and when we don't have that we go for the next best thing. Instead of just throwing our hands up in the air and exclaiming, "Gee I just don't know", we try to render some sort of educated guess. This is a necessary evil in many walks of life and game playing. In chess we have to make decisions based on heuristics, like it or not, and if we just throw our hands up and say "Gee, I just don't know what the next move should be", then we lose the game on time. If we are being chased by angry natives and we end up at the edge of a cliff with nothing but ocean water beneath us, we are forced to make a guess as to if we will survive the dive off the cliff or take a chance on trying to reason with the natives.

In the case of AI, we are making guesses, that's all we can do for now until more facts come in. You're counter points are very good though because that makes us think even deeper, and the deeper we think the better our guesses become.

Afterthought:
Wouldn't it be amazing if someday we could see a brain constructed from electromechanical parts and quantum elements? It may be much larger at first, but still act like an actual brain. If you look at some of these quantum computers around even today, they start to look almost like a biological element of some type, even if still rudimentary. As time goes on I expect all this stuff to get more and more advanced, so it will not be too much of a surprise if the first human made brain becomes a reality.
Now somewhere along the line they may even decide that it's better to power it with actual 'food' like we eat today. That would be crazy but possible. Of course, food probably isn't the best fuel but if they want to make it completely autonomous and able to reproduce, who knows what may emerge. Perhaps they can make it so that it melds in better with nature than we do right now as humans. The fuel it uses would then have to be sustainable and not interfere with other natural processes, unless we find that interference is better. Some interference is probably better, but it takes a complete and whole knowledge of the system to be able to get it right. Can we do that. Only with more advanced knowledge which i do not believe we possess right now.

 

The word quantum triggers my BS detector. Just replace magic for the word quantum. Now it sounds more scientific.
like photosynthesis and another biological process like the eye processing photons, literally anything can be explained with quantum mechanics, it tells us nothing about the human brains abilities.

 

Hello again,

Another thought...

AI can sometimes suggest 'words' that come after something you already typed. For example, if you type "For" as in the start of this sentence, it may suggest "example" which comes right after "For". That may be because the phrase "For example" comes up a lot in writing.
So what happens when AI gets so good that it KNOWS the person who is typing, and knows them good enough to know what they are going to actually say next, if that is possible. We won't have to type anything or even say anything, because the AI program would know what we were going to type or say before we even said it. So we could just lie around and let AI programs do all the talking for us. We'd never even have to open our mouths. That would be pretty weird.
Then, since AI programs can communicate via electronic signals or even WiFi or Bluetooth, we wouldn't even hear them talking or typing.
There would have to be some sort of feedback though or else our experiences would not be registered in our own minds, but perhaps at that point the AI programs would just start to take over even our own selves as each AI program acquired our own personalities.

Luckly, at least for now, this is all science fiction.

 

It's not science fiction, we have these things called executive secretaries that do that today.