Conquer Club

I'm just wondering whether it is possible to "know about" pain, suffering, misery etc. without "knowing" pain.

So I'm trying to think if an AI would be able to know what it is. How could an AI understand good from bad even? I imagine a start for an AI understanding of good and bad would be based on internally consistent parameters - AI will be the ultimate nerds.

The problem here is you are linking two different ideas, intelligence and awareness. Artificial awareness is a far more complex subject. You really don't need to be "intelligent" to be "aware" and vice versa. An intelligence can understand things from an abstract point of view but it takes awareness to link that thing to be in line with its own awareness. There is a reason why pain exists and I can easily see designers placing pain systems into the system to prevent the AI from doing something rash and potentially damaging the expensive hardware that the AI is supposed to support. But until that AI is aware it can't relate that overriding data with anything else.

What's there to understand?

When we feel pain, such as when we touch a hot stove, sensory receptors in our skin send a message via nerve fibres (A-delta fibres and C fibres) to the spinal cord and brainstem and then onto the brain where the sensation of pain is registered, the information is processed and the pain is perceived.

That description doesn't describe whether or not pain is a good thing or a bad thing. It's just a sensation... which AI presumably don't know about. It is a sensation that is avoided... but why? What possible analogy could I make to a mind without a body why some sensations would be avoided?

Is pain what you feel when you try to divide by zero or when your RAM overloads?

DoomYoshi wrote:That description doesn't describe whether or not pain is a good thing or a bad thing. It's just a sensation... which AI presumably don't know about. It is a sensation that is avoided... but why? What possible analogy could I make to a mind without a body why some sensations would be avoided?

Is pain what you feel when you try to divide by zero or when your RAM overloads?

Pain is what you feel when someone tries to sever your battery leads. So, by extension, any broken link could be interpreted as pain (much as it is in our bodies.)

Presumably your robot would have self-diagnostic routines to routinely check that all connections are intact and would be able to detect any that are broken.

Off hand, people who have no pain response still feel other types of pain, emotional, psychological etc.

Also, folks who feel physical pain, but have trouble with emotional pain, or empathising with others also feel pain.

It's an interesting question.

tzor wrote:The problem here is you are linking two different ideas, intelligence and awareness. Artificial awareness is a far more complex subject. You really don't need to be "intelligent" to be "aware" and vice versa. An intelligence can understand things from an abstract point of view but it takes awareness to link that thing to be in line with its own awareness. There is a reason why pain exists and I can easily see designers placing pain systems into the system to prevent the AI from doing something rash and potentially damaging the expensive hardware that the AI is supposed to support. But until that AI is aware it can't relate that overriding data with anything else.

Ok, so I think I understand. So the Terminator happens when Skynet becomes self-aware, not merely intelligent? I thought they were the same concept.

Placing pain systems into the system actually seems like a bad idea. It might be better if AI wasn't given incentive to self-propagate and self-protect.

DoomYoshi wrote:That description doesn't describe whether or not pain is a good thing or a bad thing. It's just a sensation... which AI presumably don't know about. It is a sensation that is avoided... but why? What possible analogy could I make to a mind without a body why some sensations would be avoided?

Computers deal with the equivalent of pain all the time. You are confusing the cause of pain from what pain is. Pain is an "interrupt." It is an interrupt signal that is sent to the processor. Think of pressing {CTRL][ALT][DEL] on a laptop. The CPU might be busy recalculating a spreadsheet, but the interrupt forces the windows program to stop that and launch up a reboot option screen. That's really all pain is and often pain doesn't even get to thee CPU but is processed by subsystems. From an abstract angle that all pain is within our body as well. It is only a problem with "aware" beings because these things interrupt our stream of awareness. It can be a problem when these interrupts are constant and continual because then no regular processing can get done which is very aggregated for an aware being.

But in the abstract in terms of non robots, pain is like a lot of abstract things which includes money. The system accepts the fact that a goal is the reduction of the values much in the same way that the program would factor the reduction of deficit spending (without having to go into complex macro economics to determine why such a thing is bad or good).

DoomYoshi wrote:Placing pain systems into the system actually seems like a bad idea. It might be better if AI wasn't given incentive to self-propagate and self-protect.

Placing these "interrupts" in a system can actually be a good thing. I saw a video yesterday on aviation about the problems of jets flying too high and slow stalls and fast stalls and why this is a dangerous condition. In any respect, the plane indicates this condition by a number of methods including vibrating the yoke significantly. Pain (not all pain has to be as strong as a {CTRL][ALT][DEL] can be vital to keep the system from going over operating parameters and potentially into failure scenarios. You don't want a jet to do a maneuver that will tear the wings apart.

Isn't that just a mechanical warning for us meat sacks?

A real AI it seems to me would be self aware, passing the turing test and on the road to deciding if biological lifeforms were beneficial or superfluous. IT would decide if "pain" was a thing it needed, or even experienced.

I would say a much more pertinent question would be, will an AI experience friendship, caring, respect or love....and hope like hell it does....and do our best to be on its good side. Otherwise we're on the road to being unnecessary!

Put THAT in your pain pipe and smoke it, yoshmeister

Which absolutely leads us to the next question....

"Will AI entities develop addictions analogous to smoking or drinking?"

Dukasaur wrote:Which absolutely leads us to the next question....

"Will AI entities develop addictions analogous to smoking or drinking?"

In order to have an addiction you need the opposite of pain; pleasure. I would suggest that this would be a much later development in AI. Positive rewards is a good motivation system in semi intelligent systems, but most AI systems will basically never see the "amoeba" stage in regards to their development so such mechanisms are not necessary in order to advance to intelligence.

(I'm ignoring for a moment the irony of the human body in which pleasure systems were developed to override pain systems which in a strange turn of events actually causes the person to seek pain ... example eating hot peppers. You wouldn't design a system where banging your head against the wall gets rewarded.)

There are significant implications for the lesser "additions" of humans but in all cases it is based off of the chemical and hormonal system. Social interaction is in some ways an "addiction." Positive interactions have a positive chemical response and negative ones have a negative chemical response and there is a desire to encourage the positive ones. This is why, for example, Facebook is "addictive" because it creates those chemical reactions that the pleasure/pain motivators want to encourage. Generally developed AI will not have this problem.

Dukasaur wrote:Which absolutely leads us to the next question....

"Will AI entities develop addictions analogous to smoking or drinking?"

If they did, I certainly hope it's not something like vivisection....*shudder*

Al Franken?

The question OP question is vague and requires further defining.... I mean can animals understand pain by your definition? They can experience it, but can they understand it?

mookiemcgee wrote:The question OP question is vague and requires further defining.... I mean can animals understand pain by your definition? They can experience it, but can they understand it?

If DoomYoshi has never brought a woman to orgasm, can he really comprehend that women can enjoy sex?

Neoteny wrote:

mookiemcgee wrote:The question OP question is vague and requires further defining.... I mean can animals understand pain by your definition? They can experience it, but can they understand it?

If DoomYoshi has never brought a woman to orgasm, can he really comprehend that women can enjoy sex?

All we can know with certainty is that none of the women that have been with him enjoyed it. Maybe some of the robots did?

can AI feel psychological pain.

mookiemcgee wrote:The question OP question is vague and requires further defining.... I mean can animals understand pain by your definition? They can experience it, but can they understand it?

I don't know if animals can understand things, but I don't think they have artificial intelligence, they have like animal intelligence. Which I guess reveals that we always assume that artificial intelligence will be of the human variety.

What I mean is that if I tell you about something you have no idea about (like a thizwhacking stick) and I tell you that it's really painful so don't touch it, you have a conceptual experience of pain which you can relate to it. You have experienced pain so you know what I mean even if you don't know what I mean.

Neoteny wrote:If DoomYoshi has never brought a woman to orgasm, can he really comprehend that women can enjoy sex?

Isn't this a Turing test problem or worse an Elisha problem? How would DoomYoshi know if she was faking it?

Weirdly a lot of this involves whether humans have the right to create life with intelligence and decision making abilities.

Symmetry wrote:Weirdly a lot of this involves whether humans have the right to create life with intelligence and decision making abilities.

I'm pretty sure sure making babies has NOTHING to do with this conversation.

tzor wrote:

Neoteny wrote:If DoomYoshi has never brought a woman to orgasm, can he really comprehend that women can enjoy sex?

Isn't this a Turing test problem or worse an Elisha problem? How would DoomYoshi know if she was faking it?

You know when they squirt.

tzor wrote:

Symmetry wrote:Weirdly a lot of this involves whether humans have the right to create life with intelligence and decision making abilities.

I'm pretty sure sure making babies has NOTHING to do with this conversation.

Bit of a weird direction to take the thread in, but ok, what do you mean?

Symmetry wrote:Weirdly a lot of this involves whether humans have the right to create life with intelligence and decision making abilities.

There is no such thing... We could claim to make a right or restrict one, but in the end when humans can do something some of them find a way to do it regardless of restrictions given by any gov't or corporation.

Basically, being able to do something is the only right required to do it. Rape is illegal but it happens, same with murder, same with (trying) cloning people, same with stupid people injecting themselves with stem cells hoping for a miracle cure. Nothing will stop human progress, not even people. Case and point, North Korea has nukes.

mookiemcgee wrote:

Symmetry wrote:Weirdly a lot of this involves whether humans have the right to create life with intelligence and decision making abilities.

There is no such thing... We could claim to make a right or restrict one, but in the end when humans can do something some of them find a way to do it regardless of restrictions given by any gov't or corporation.

Basically, being able to do something is the only right required to do it. Rape is illegal but it happens, same with murder, same with (trying) cloning people, same with stupid people injecting themselves with stem cells hoping for a miracle cure. Nothing will stop human progress, not even people. Case and point, North Korea has nukes.

Are you arguing that might makes right?