Do robots have feelings? [Archive]

View Full Version : Do robots have feelings?

Mike

26-12-2005, 21:22

I was thinking today, and all humans are are really complex robots. We are "programmed" to feel emotions, via our sensors. So, if I program a robot to say "Ow!" when you punch it, do you feel bad? Where do we draw the line between robots and humans? Theres a lot more questions I want to ask, but I can't really get them out right now... but it basically comes down to

When do robots become "living"?

greencactus3

26-12-2005, 21:49

sciencenerd

26-12-2005, 21:54

Mike

26-12-2005, 21:55

wow...... you are BORED. :p

i will hafta say robots are "living" once they stop taking orders from idiots. like us.
Heh, "bored" to some... a thinker to others :)

So if I program a robot

pwm01 = 255;
pwm02 = 255;

It will go full forward, not caring how much you play on the joysticks. Is this robot now living?

Mike

26-12-2005, 21:57

I would probably say that you could call a robot living once it gets complex enough that it is impossible to determine how a robot will react to something (complex enough, mind you, not random enough) you could call it sentient.
Ahh, I thought I had you... but then I read the "not random enough" part. Smart.

So, if you throw a baseball at my face, I am "programmed" to duck out of the way or try and catch it (depending on my confidence in my baseball skills :P). You can determine that I will duck when you throw a large blunt object at me, does that mean I am not living?

Andrew Blair

26-12-2005, 22:01

So, if you throw a baseball at my face, I am "programmed" to duck out of the way or try and catch it (depending on my confidence in my baseball skills :P). You can determine that I will duck when you throw a large blunt object at me, does that mean I am not living?

Aha, I assume that you will duck. You probably will. But the question is, if you escape the large, blunt object, will you begin ruthlessly beating me with a baseball bat, or will you walk away, singing some nondescript show tune like nothing happened?

Mike

26-12-2005, 22:14

Aha, I assume that you will duck. You probably will. But the question is, if you escape the large, blunt object, will you begin ruthlessly beating me with a baseball bat, or will you walk away, singing some nondescript show tune like nothing happened?
Well, that depends on many things. Mainly my personality. But, using some fuzzy logic, you can give a robot personality traits as well. I know very little about fuzzy logic, so I'm going to use a discrete value between 0 and 100 for now.

int angry = 63;
int compassionate = 25;
int nice = 17;
int aggressive = 97;

So there you have my personality. I'm an somewhat easily angered guy, that will become very aggressive once I am angered. From this data you can assume that I will pick up that bat and beat you.

int angry = 27;
int compassionate = 73;
int nice = 82;
int aggressive = 12;

From this data, however, you can assume that I will probably ask you why you would do such a thing and then probably forgive you.

Now, I'm not expert in the workings of emotions, but I'm assuming this is kinda-sorta how a human works. These traits are variable, depending on our environment, upbringing, and an infinite amount of other things.

Now, I kinda forgot where I was going with this, so bear with me here. I think I was trying to make a link between how humans react to things according to their upbringing, environment, and other factors and how robots can be programmed to react also depending on their environment and other factors.

sciencenerd

26-12-2005, 22:17

Ahh, I thought I had you... but then I read the "not random enough" part. Smart.

So, if you throw a baseball at my face, I am "programmed" to duck out of the way or try and catch it (depending on my confidence in my baseball skills :P). You can determine that I will duck when you throw a large blunt object at me, does that mean I am not living?

Well in this case, you only had two best options for your own good. You analyze both of them, and choose whichever one seems to be best. That's not predictability, that's intelligence.

Okay, I'm just brainstorming here, but you made me think of a new definition. Maybe this one's a little better.

When a robot can be presented with a situation that it has had no precedent of (or programming about), and, from the information it can gather, choose the best way to react to it, then it is sentient.

For example: If take a "sentient" robot that knows how to climb on objects and push on objects. Now, if you place it in a room with a high shelf (that you have programmed it to try to get to) and a stool on the opposite side of the room, it would be able to figure out how to push the stool over, climb on it and reach the shelf.

Of course, all of this is getting a little off the topic of "feelings". The question is, how do you define feelings? Feelings are a way for your body to express a desire. In that respect I would say yes, robots do have "feelings".

Mike

26-12-2005, 22:35

When a robot can be presented with a situation that it has had no precedent of (or programming about), and, from the information it can gather, choose the best way to react to it, then it is sentient.

Ahh, now we are getting somewhere. I can't answer this with a little snippet of code. But rather this leads into another question. Are humans very complex programs? I say that we are programmed, but programmed to be learning creatures. If you were to place a baby in this situation, it would not know how to get the candy on the top shelf. However, place that same baby (now more of a toddler) in the room a few years later and I believe it would now know how to get that candy.

So, if I program a robot to learn, is it living? Programming a robot to learn is no easy task, to say the least. One (very very very inefficient idea) that I have is to approach it with a brute force attack, of sorts. Try every possible combination of using the chair until you find the one that works, then log that technique for use later.

KenWittlief

27-12-2005, 10:42

you are getting into two areas: feelings and living.

First you have to define what feelings are, and what it means to be alive.

Robots and computers have already passed several significant milestones:

1. Beat the world champion at chess.
2. Killed people (I think the first was a car assembly line robot in Japan, worker was crushed by a robot)
3. Reproduction: computer viruses can self replicated.

a good indication of feelings will be when a robot gets punched and:

1. says "ow!"
2. then says "that really stings you jerk!"
3. then refuses to do anything until you apologize :^)

lukevanoort

27-12-2005, 11:40

The excellent book, The Age of Spiritual Machines by Ray Kurzwell deals with this exact question, and is quite an interesting read.

Mike

27-12-2005, 13:07

The excellent book, The Age of Spiritual Machines by Ray Kurzwell deals with this exact question, and is quite an interesting read.
Thanks for the tip. I just looked it up on Amazon, and it's pretty cheap ($10). Next time I'm near my Barnes&Noble I'll check it out.

John Gutmann

01-01-2006, 01:19

Ahh, now we are getting somewhere. I can't answer this with a little snippet of code. But rather this leads into another question. Are humans very complex programs? I say that we are programmed, but programmed to be learning creatures. If you were to place a baby in this situation, it would not know how to get the candy on the top shelf. However, place that same baby (now more of a toddler) in the room a few years later and I believe it would now know how to get that candy.

So, if I program a robot to learn, is it living? Programming a robot to learn is no easy task, to say the least. One (very very very inefficient idea) that I have is to approach it with a brute force attack, of sorts. Try every possible combination of using the chair until you find the one that works, then log that technique for use later.

Being able to learn is one trait that makes something living. aside with being able to learn is being able to adapt to your evironment. Just because a robot doesn't take the easiest root doesn't make it not living. If it takes the smart(intelligent) path it doesn't make it living either. I know many stupid people. Many many stupid people. But that doesn't mean they aren't alive. Something that you can can say defines living things is the ability to adapt, as I stated earlier.

For example Animals adapt to cold by growing fur(multiple layers). They adapt by living in packs. they adapt to water by growing fins and the shapes of their bodies.

-Plants adapt by growing towards the sun.

But everything in a way is kinda pre-programmed to an extent.

For example:
One of the biggest ones I can think of right now is Fibonacci's sequece.
The mating of rabbits. you start wth one rabbit then you add another(male + female)and the pattern is 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233......

the same can be said about the squirly seashells, pine cone tines, branches on a bush, everywhere you look.
Is that squence of #'s programed into everything or is it just a coincidence?

But I am sure you were talking about how a robot acts. I would say one important point is that living things can physically change is almost any way that is predictable to some extent. To be able to have a living robot in pure programming I think is impossible Just the simple logic of a robot to beable to learn is complex not to even start to think about how to program a memory like our that is photographic and our brains store everything we expirience(it is just a matter of if we can remember it).

In my honest opinion man kind can never and will never create a "living robot" because the create somethign to the extent is like an infinite project, even now we are learning more about living things.

Ashley Christine

01-01-2006, 14:49

I hug my robot, its cute.
As for are they living... I dont know what to say. FIRST robots arent living obviously.
Mike, dont worry, you arent a robot, or are you?! and if you are, who built you? and who is that all-mighty programmer? [You should meet up with that guy!]

KenWittlief

01-01-2006, 15:14

Feelings seems like the easy part. In a way dont we have stuff now that tells us how it feels?

My car tells me when it doesnt 'feel good' by flashing a 'SERVICE ENGINE SOON' light on the dashboard, and when its feeling low on energy by blinking the 'LOW FUEL' light

my PDA and laptop tell me when they are feeling run down and need to be plugged in to recharged.

Life as we know it is a little more challenging. Things that are living can:

1. generate their own fuel or find it
2. self repair
3. self replicate (with error detection and correction).

Robots like the Roomba can find a wall plug or their charging station when they are low on energy, thats similar to #1 on the list. Now picture a Roomba like device that can repair itself when it falls down the stairs, and can build a duplicate of itself when it realizes it cannot keepup with its work all by itself.

Cody Carey

01-01-2006, 15:37

I'll consider a robot to be living when it won't let me turn it off.

KenWittlief

01-01-2006, 16:07

I'll consider a robot to be living when it won't let me turn it off.

you mean like most PCs running windows 95? :^)

Ashley Christine

01-01-2006, 17:58

Once a robot can reproduce then it will be living. LOL.

Mike

01-01-2006, 18:47

1. generate their own fuel or find it
You stated the Roomba, no need for me to reiterate.
2. self repair
Hmm. I can't really think of a popular/in production robot that does this now, but I can't picture this being anything beyond a somewhat complex self-debugging system. What are the logical things you do when your finger is hurting you? Rest it. Splint it maybe. Ice it. Etc. The robot would be programmed to recognize a set of symptoms and treatments.
Once a robot can reproduce then it will be living. LOL.
3. self replicate (with error detection and correction).

http://science.slashdot.org/hardware/05/02/02/2253245.shtml?tid=216&tid=14

The Guardian unlimited is reporting that Korean roboticist Kim Jong-Hwan, who founded the robot football (soccer) World Cup, and is the director of the ITRC-Intelligent Robot Research Centre, has developed a series of artificial chromosomes that, he says, will allow robots to feel lusty, and could eventually lead to them reproducing.

http://hardware.slashdot.org/article.pl?sid=05/05/11/2242239

Scientists at Cornell University have created small robots that can build copies of themselves.

As far as "with error detection and correction", I'm not totally sure what you mean by that. What if this robot is programmed with a PID loop of sorts, and when it reproduces the PID is tweaked because it notices that it needs to move slower/etc.

I'll consider a robot to be living when it won't let me turn it off.
Well, instead of that big red switch on the robot, just directly connect the wires. Yes, you could just cut the wires, which are the lifeline of the robot, but you could also cut a living beings lifeline (blood vessels, windpipe, etc.)

As far as sparksandtabs post on adaptation... that is a tough one. Now, if I understand evolution correctly, animals naturally do not adapt to their environment. Instead, an animal is born with a genetic mutation that allows it to survive in a new area. This animal out-lives and out-reproduces all the other "normal" gened animals, eventually becoming the predominant gene type. So now this ties in with the topic of robotic reproduction. Going back to the PID example, would a robot with the better PID tuning not out-live the other robots, with an inferior tuning? This robot will recognize that it does not need to change that aspect of itself, and pass that aspect down to it's children. The children will then find an unsatisfactory aspect of it's programming, fix it and pass it down to their children, so on and so forth.

This post may be confusing and unorganized, but it's hard to organize thoughts like this :o

KenWittlief

02-01-2006, 01:06

Our DNA contains duplicate (redundant) information, which allows reproduction to take place with error detection and correction.

An engineering equivalent is flight control systems, which have 3 or 4 copies of each unit, that constantly cross check each other. If one is found to be operating differently than the rest then it is locked out by the others. Thats error detection.

Error correction (fixing the failed component in real time) would be a little more complex. Our bodies fix themselves without our conscious involvement or awareness. Our blood clots to stop leaks, then our skin, muscle and bones repair themselves from the basic elements present in our blood (from the food we have eaten).

A robotic equivalent would be a fluid that contains iron, aluminum, copper, semiconductors... and nanotechnology that is able to detect when part of the robot has failed or been damaged, and pulls the needed molecules from that fluid, and rebuilds the damaged component molecule by molecule.

Thats the kind of self repair you would need to approach robotic life.

greencactus3

02-01-2006, 01:21

well theres those new plastics with "self repair". sumthing about catalysts in them. i forgot

sciencenerd

02-01-2006, 01:31

[a little off topic] A very cool program that replicates selection in a population is gene pool (http://www.swimbots.com/). I let it run a few days once, and the swimbots that came out were amazingly well developed.

lukevanoort

02-01-2006, 19:04

Saru29

02-01-2006, 19:48

When a robot army comes marching in to force us into body pods put us ina virtual world and lets call this the matrix for some random reason and in the matrix some dude can dodge bullets comes and talks to the machines while being blind...Then i will believe robots are alive.