Ooh boy, a robot sentience discussion
Quote: "dad: "well that's the difference between humans and machines. we have a mind and can do more than just calculations" "
(This idea of what a computer *really* is is taken from Douglas Adams.) Once upon a time, a man built a computer to do sums. So we thought a computer was a calculator. Then we saw that all those on-off switches could be more complex and could stand for symbols, like letters - and we thought that it was a typewriter. People started using it for other things like playing video games and we thought it was a gaming machine.
But what a computer *actually* is, is a simulation engine. All those on-off switches can be ordered in such a way as to simulate pretty much anything. So when a computer can do so many calculations that it can solve puzzles like humans and act like a human and think like a human (or appear to), it is a simulation of a human being.
But, if the simulation does everything the original does, then is it truly a "fake"? If you make a copy of a file on your computer, it is no less "real" or "fake" then the original file due to the properties of digital media. Is there a point where the fake becomes the real?
Quote: "mum: "but if the robot says 'hello' then it says it because that's what you told it to""
To put forward my view on this I'm going to have to explain a rather complex idea.
1) You have a brain, and this brain thinks. That is, your brain takes information and makes a judgement on that information based on your past experiences. If, for example, you had a choice of pizza or salad the brain would look in your memory, see that you don't like salad and use that information to make a choice - ie. pizza.
2) Let's say you have a man in a room. You have complete control over his memories and over his physical wellbeing. Every day, you wipe his memories of the previous day, and return his physical wellbeing to a pre-determined default state. Mentally and physically, he is "reset", so to speak.
3) As long as all external stimuli are identical from day to day, this man will always behave in the same way because he will have the same external information and the same memories. He would even think the same thoughts as in all previous days, because the brain would be receiving the same stimuli day after day after day, and would give out the same results.
4) There is no such thing as free will - we only do what our brains end up deciding to do. We are, as it were, slaves to causality. (But we should never behave as though we knew this, because then nobody would get anything done.)
I'm not sure if anybody understood that - certainly very few people I've met have agreed with me on this, so I may be wrong. But if this is the case, then we ARE calculators. We DO say "hello" because we have been programmed to do so - albeit by evolution and social contact. I believe the question should not be "what makes a robot's will genuine?" but "what makes OUR will genuine?"
Also, on the subject of emotions. Firstly I'd like to disassociate emotion from sentience. I haven't studied this so I could well be wrong - however, some people (who I sympathise with immensely) have, in the past, been lobotomised. This is when a section of the brain is removed which renders them incapable of emotion. If these people are still the be considered sentient (and I believe that they are, since no other higher brain functions have been removed) then we must also consider the possibility that a robot could be sentient but not necessarily capable of emotion.
Finally, the "truthfulness" of emotions. Many (and, I believe, all) emotions are a result of evolution: fear makes us run from animals, anger makes us more dangerous, and friendship and love help to create a mutually beneficial bond between animals/people. All these help to ensure the survival of the genes that produce them, which will (ideally) be passed on to the next generation. (This is the so-called "selfish gene".) Emotions are able to make us feel various things via chemicals released into the brain.
Now, what if a robot were, for example, programmed with an "instinctive" fear emotion? It would need to recognise certain triggers (a relatively simple matter of shape recognition to detect, say, snakes and spiders) and it would need some way of releasing the emotion into the brain once triggered. In humans this is chemicals; in the robot this depends on how the brain works, but should still be possible since all that's needed is a sort of "gosub" command, but for the brain. Now, I admit that this all looks rather phony: our poor robot-building scientist has had to go to great lengths, in this example, to create emotion in this robot. But the fact remains that the robot reacts to certain things (eg. snakes and spiders) in the same way as a human,
and presumably undergoes the same feelings of terror as a human being would under similar circumstances because we are directly manipulating its brain, which is where it feels things (and where humans feel things, emotions included). Although the robot is undergoing a "fake" emotion, it believes the emotion to be real - and, after all, what is emotion? Chemicals released into the brain? Why, then, human emotions are also fake, by the same token.
