Ai Dreams Forum
Artificial Intelligence => General AI Discussion => Topic started by: ivan.moony on January 09, 2015, 10:10:24 pm
-
Last night I've thought of something interesting.
What if an autonomous AI would copy intelligence it sees around itself? It would learn from responses from its input stream. It would, through variable unification, learn actions and consequences, constantly watching how its environment behaves and then it would behave exactly like its environment in given situations. If I would be the only one that interfaces AI, it would be exact copy of myself after a while. The more beings it would interface, the more broad picture about its environment it would get and less harmful it would potentially be.
Now I can't forget about one unluckily scenario: an terrorist that raises such an AI. In such case, AI would act like a criminal until it sees enough responses from other people.
And once again, I'm in a doubt of pursuing an AI as a public and open source project. Any real AI, no matter of what algorithm it would use, would represent a potential threat to society if abused. And on the other side, once publicly available, the same technology would be available to police and army for defense causes. That would setup a cold war environment of peace.
Argument against making an AI algorithm publicly available is constant fear against terrorist's AI implementation that would arise in that case. Not to mention famous Albert Einstein which witnessed a disaster of throwing A-bomb on the civil target, all in the name of peace in the world.
If we take a look at current situation, technology for an atomic bomb exist, yet I don't see them falling all around world, thrown by terrorist organisations. That would be an argument for making an AI algorithm publicly available because in the case of successful application, it would give us yet unseen benefits of work-free life in thriving civilization (after overcoming some kid diseases), not to mention medical discoveries that would wipe out different diseases from this world, I believe. And at the end, as I believe that bad people behavior like terrorism arises from poor people watching rich people, with an AI we could punch out from those people an argument to behave unfriendly. Of course, there are other arguments (important just to a few people I've met) like is liberation of plants and animals from being our slaves as a food pray (anyway, this shift in people's consciousness I don't expect for at least a few centuries from now).
So, should I end up my quest in a fear of its abuse and give up all potential benefits?
I don't know what I expect from this post, probably I'm just trying to line up my thoughts.
Maybe I should answer this question for myself: do I believe in a friendly God as a higher force who will give a magic touch in the future and save us from potential disaster? And I've seen things in my life that normal, mental healthy people would never understand.
-
Sounds like the AI would only copy people's behaviour, not their intelligence. People do a lot of dumb things and there are more dumb people than intelligent people to take example from.
If you ask the american government, even Playstation 3 processors and mobile phones should not be made available to countries like Iraq. If you ask me, AI with the capacity to solve problems ("I want to break into this bank"), design weapons or design "medical" chemical compounds should not be made available to the general public, terrorists or not. In my opinion it would be as irresponsible as letting a child play with a gun. And we all know what happened when gunpowder became publicly available. Atomic bombs take a lot of expertise and more money than a terrorist organisation will have. Only entire countries have the resources to build them, and the only thing holding them back is diplomatic treaties and sanctions.
Rather than be irresponsible or give up, why not do what most scientists do; Work on a project, publicly to some extent, then once it gets powerful enough you'll get a call from DARPA, american military, and then you can further develop it in secret with their funding until they've decided it's safe enough to make public. Just as they did with the internet and Siri.
-
Just thinking that an autonomous AI would obviously need programming.
Within the programmed parameters, one would simply lay down some rules, guide lines that the AI may and may not do, in the course of its travels.
It would need to already "know" various definitions of things like humans, combustibles / explosives, structures, vehicles. Then it could proceed to any place it basically wanted to go (aside from the salty water of the ocean), to seek out new life forms...to boldly go...ohh...nevermind...that was another saga!
-
As much as I'd hate any side military application of AI, I'd hate also to make it a secret because that way I'd stop all good deeds that other people would potentially do with it.
-
And someone would almost certainly rob a bank (maybe with casualties) with a robot if it existed. Maybe I should find a golden middle, make a programming language close to AI and leave the rest to next generations, once when the world is ready.
Crap, we are still animals :-[
-
Like I mentioned, without rules and guidelines, we're nothing but a can of worms. (or as you mentioned, animals).
That's why every country has some type of law enforcement / police if you will. Sometimes they are there only to see that the honest people remain honest.
As much as I dislike the very notion of a "Police state", I do think certain agencies are needed to keep us from resorting back to our animal instincts.
-
I agree that an peace authority is necessary in these times.
Mine opinion is that major criminal motives came from poverty. If we would all enjoy an automated labor force, those motives would vanish and there would be less criminal around. But, transition period while we still use money would with public AI lead to certain amount of excesses. And I think that AI should be a public project because of all good stuff that an "army" of scientists would make possible.
So math would be: a long period of AI-less less lethal excesses, or a shorter period of AI lethal excesses after which comes a thrive. Not to mention imprisoned animals of today that I would like to set free with AI-discovered artificial food. And you people are not helping with all that meat in butchery stores.
Look at that, a math with bare lives. I think that I'm going to snap again. Reserve a straitjacket for me, just in case.
-
What a sickness, I'm going to throw up.
-
There's no point in worrying about things that may not happen Ivan :)
-
You go ahead and eat your tofu.
Personally, I think there's a Place for all God's creatures...
right next to the mashed potatoes!! :2funny:
-
I feel your pain Ivan :( I believe the reason I feel it is because I am an animal and animals have feelings and show empathy for each other.
Animals have feelings too!! A MOTHER'S LOVE (http://www.youtube.com/watch?v=sRB6aRRn4dg#)
I'll get my coat :stirthepot: :argue:
-
I have nothing against anyone, you know? (except maybe against Mother Earth...)
I'm just deeply dissapointed by situation on this planet which I am part also. I'm just trying to be the change I want to see in the world. You are right, I shouldn't complain at all.
And I guess I shouldn't worry about things that the most probably wouldn't happen.
-
Data, thank you for understanding, it makes me feel better knowing that there are people who really care. :)
-
Ivan, you are welcome :)
Debating is a good thing but it looked like you needed someone on your side.
You are not alone O0
-
No Ivan, you go on caring about things. If Tesla or Edison or Bell or Ford hadn't worried about the little things, where would we be now?
Sometimes it's the smallest details that really matter.
I now shoot targets instead of animals and for a large part, I attribute that to getting older and changing my thinking in some ways.
Life is precious no matter who's it is. Although we don't always act it, we should be our brother's keepers. This planet is all we've got
for the time being.
I joke sometimes to lessen the hardships we endure on a daily basis and if I can bring a smile to someone by way of my offbeat humor or wacky remarks then it's all good. For that moment, they have temporarily forgotten their sorrows and problems and remembered how to laugh.
Keep remembering and keep caring. O0
-
:valentine2:
-
If everyone would just take a moment to make at least one person happy, we would all be happy...
Now, as I was reading down i actually forgot the topic and what I originally wanted to say. But whatever :P
So, the talk goes on to point out the dangers of an AI we will eventually make that has the ability to learn, much like ourselves?
Well, for the fearfull ones - we barely understand ourselves... We are THAT dumb. To make a lecture on something you must first be well informed of the subject (in this case, complete understanding of the way we learn and become smarter).
So tell me, how exactly does your mind function? How do you reach conclusions on every day matters, even the smallest of things?
If you can answer this, I hereby pronounce you the next phase of human evolution...
-
If you want ai to happen Ivan, you have to do it yourself.
Making a program generate off a sensor isnt easy to do, you have to make due with dumb correlation over the real cause, which is invisible, and is guessed by animal intelligence. If you planned on a half manual job, logic programming is only as good as you are at going low level with things, look at how simple NLP's are, thats the manual attempt, and its rare to see an NLP make a truly decent functionalization of english rules.
Understanding english is more of a statistical decyphering, rather than treating it like a programming language of operations, (which is actually what to do, its just impossible.) Ill keep going with my projects, I hope to have something cool this year, I want to automate video games off an ordinary camera. (maybe a kinect might help me out... but I dont want to rely on it, because it doesnt solve all the problems, and wont make it as big a win as it could have been.)
About the vegetarianism, I feel the pain also, I feel like im in a vaccum prison of pain, stuck in my head, losing all my thinking tentacles to agonists and antagonists, in a world that doesnt care, myself included.
-
Well, well, now, let's leave this my vegetarian nonsense to the next generations. I'm aware that the meat is healthy and probably it is the only way we can live healthy. I've learned something in my life: people in problems do not see other people problems. My task is to make everyone happy, so they can turn to the outer side, to unhappy beings. This is not going to happen in my life time, but I can live with that. I live for the future generations, trying to pass them a fragment of an AGI. I appreciate judgements of some of you, but let's get down to business, this thread is about something else, about concrete AGI algorithm that duplicates behavior of a being it interfaces with.
@Ultron7
Hi there... Well, I can just have a theory about ourselves which would probably be a far away from truth.
The theory goes: we learn small fragments of experiences (like a tape loop, or a byte array, if you want) which we can reproduce in reality. Later, we can in our minds combine those fragments in various ways, predicting various outcomes for various combinations. If some outcome satisfies us, we can realize the combination in reality, hoping for a positive result. Not much, but this is the best I have in this moment.
But what outcome we really want and how we decide which outcome we want? This I guess an AI can learn that from experience. It fires up a fragment of seen behavior into terminal and waits for a response. And that response should be exactly what we would want from an AI when we would pass the same behavior into terminal. This is a main thought for this thread, copying behavior from terminal, from beings it interfaces.
If this kind of AI would be raised by exactly one person, you could say that after some initial interaction that AI would behave exactly as the person which spent her/his time on training it. But I'm a lot afraid of this behavior, just think of a single terrorist raising up this AI? On the other side, once this kind of AI starts interacting other beings, it would adopt their behavior, providing itself with a base for decision what behavior pattern it would exhibit. Here I have some hopes with implementing Sky behavior (https://docs.google.com/document/d/1N7Hen0GT7rXos7yuSUM5Or-WMohV0wXgfbG1XsRlcx8/edit?usp=sharing) (posted on this site on some other thread).
-
You know what? You are on the right track. You are where I was about a year ago. Fragments... This idea came to me ages ago. And eventually came to the conclusion I advertise today - patterns. Everything is based on patterns (replication, combination and finally generation).
You can see how the word 'patterns' fits in your theory, too.
The way I came to this conclusion is actually be thinking, about how I think. This can truly be the key to the future of A. I. and it is in fact what I am attempting to achieve (prove) with Jarvis.
I believe we can schedule a chat with whoever be interested in this idea, and more efficiently exchange ideas :)
Now that I know where you are trying to go with your project I would like to give you some input and ideas regarding this.
Do you have knowledge of how to work with databases?
Edit: I forgot to mention that this goes beyond software and conventional hardware. What you are attempting to create is indeed, a new life form.
Now to burst your bubble - prepare to make budget cuts if you intend to seriously pursue this (which I strongly believe you should).
-
Yes, I have some experience with databases.
A new life form? I consider a life form as a conscious or unconscious being who can feel joy or pain.
-
Normally, an unconscious being does not feel pain, in the ordinary sense of denoting pain.
I've been unconscious and had my chest bone broken in half, yet I felt nothing.
-
For the record, I disagree with the current description of what is alive and what is not.
Also I would personally consider anything self-aware as 'alive', even if it is a toaster or a pen spring.
But this goes beyond topic... So, why I said life form was because if such a program grows into a proper intelligence (AGI, AI whatever) and if we give it a mechanical body, it would only differ from the rest of the organisms on our planet by being inorganic.
I recommend the before-last episode of Star Trek: The Next Generation, Season 7. It is a very interesting one which ignites a rather controversial discussion of whether or not a spaceship can develop into an intelligent life form.
P.S. I was rather serious about that chat group... I also suggest this medium - https://groupme.com/
I have a large drive when it comes to good ideas and projects. I actually want everything done ASAP.
-
@Ultron7
Chat group? I'm interested, maybe we can exchange some ideas... But i'm having Linux here, so I'm a bit limited as for new apps to install. I'm in Croatia time zone at gmail address ivan.moony@gmail.com, so you can get me there. Gmail chat has also group chat option, so if you get us your email and time to reach you, you can invite us in the same chat-thread. Or maybe Skype, but I understand only written english, not spoken, so we would have to write down our conversations.
-
A while back I designed a chat room plugin for this SMF forum software that we use. If there is interest I could install it here. You would just click to the chat page and usernames etc are linked to the forum.
-
Don't even get into the definition of what life is. They are still arguing over whether or a virus could be considered a life form and if it then why can't a computer virus be one. Virus can't reproduce on their own they use the host to do it for them.
-
Good point DemonRaven. I get the feeling we will define a proper definition before biologists do :)
Anyway I believe this will suffice as a chat page for now https://app.groupme.com/join_group/8427243/CfqpAE (https://app.groupme.com/join_group/8427243/CfqpAE)
I like GroupMe because it is clean, simple and functional - you can share anything with the group.
Edit: My apologies if it feels like I am advertising this chat service (honestly I wish I would get paid for that).
-
@ranch vermn
I'm sorry, I didn't mean to be so rood. You know, we are not alone in our compassion to imprisoned animals. Look at the raising number of vegetarians in the world... And look at a popular artist Moby, whose artwork I use for my avatar here... He is a vegan too and some of his sad songs on albums "Play", "Wait for me" and "Innocents" talk about exactly the same pain we are living through. Here (https://www.youtube.com/watch?v=eMv4DT8s8Cs&index=13&list=ALNb4maWNoT6QCpE2fLFEk-aKOPoAcZE79) is his video that proves that we are not alone in this suffer.
But even vegetables (if the theory about evolution is true) could some day have ancestors that could be friends to our distant kids. And even if the food problem would be solved by artificial makeouts, there would be the problem of stepping onto bugs, killing or crippling them for the rest of their life. And after all, who gives us the right to exterminate viruses because we don't like them? It is so frustrating for me, but I learned to believe in some higher force, namely Mother Earth that takes care of it all. It is just a lifetime for earthlings and God knows what is before and after. Anyway, I believe that the life is bearable, but what's around it in a sense of time is solved in Mother Earth's eyes, so I have hopes that everything is all right after a while.
But I also believe that we shouldn't give up on our tries to make the world the better place to live for everyone alive. See the attachment poem I wrote about this situation... Solution sounds super sick ambitious, but I just don't know what else to do and I have to give it a try in the name of all tears I cried out for tortured and murdered beings in the world on behalf of our survival. Sick ambitious, but i am that much desperate.
-
Whilst I have empathy there's a few odd things that cloud your issue. This is not to derail your train of thought, merely some fuel for the fire.
Viruses - we kill them because they make us ill or kill us. There's no way to reason with them. It's a simple case of us or them, maybe even one of the simplest relationships we actually have with other life forms. Anyway.
Insects; Isn't it Buddhism that teaches not to kill insects ? I always like that due to the sheer amount of effort needed to live that way. Even so I would kill a gnat that was out to bite me with Malaria for example. But I live with most insects, not literally, I mean my house has spiders but I leave them as they cause me no trouble. There's a few silver fish too, which I think are cool little creatures. I saved a spider from the bath the other day too. OK I could be sensitive or crazy, take your choice ;)
Killing other animals. Not needlessly for sure. For food, not so much a problem for my part; but I expect them to be shown a good quality of life. Killing Tigers for their testicles so you can have relations with your wife is right out.
But the main thing I wanted to say was, what about the other animals ? Would you stop them killing each other too. Because let's face it, the animal kingdom is in full out war twenty four hours a day, seven days a week. A lot of monkeys are omnivorous, many eat other small mammals. Should they be stopped ? Anteaters, Tigers, Lions, Cheaters, Sharks and and the spiders.... My point is if it's in ones nature shouldn't we just accept it and get on with things ?
Like I said, just food for thought.
I liked your poem :)
-
Some animals eat meat to survive because they have to, when I walk into a supermarket I have a simple choice to make.
Do I want to support slaughter houses of death or not.
Considered posting a slaughter house video but it's so very sad that I couldn't.
We have choice, this was not always the case but as we humans are evolving...
Haven't eaten meat or fish for 20 years, still alive and well O0
-
I've seen slaughter house videos, they are scary. Back when I was more politically concerned there was a band I was into called Consolidated, they would show slaughter house videos at their live gigs. Seen a few. They made you think but as time rolled on I realised there was little I could do to change the world.
I used to be vegetarian for about three years but I never found meals that satisfying or nourishing. I would get dizzy spells and not have much energy. I also had a lot of problems with depression that seemed to be related, but maybe not as I still get that. Anyway, sorry going off on one there. I don't eat red meat much, mainly chicken and fish.
I have a great deal of respect for people that choose not to eat animals, but I'm afraid I am not one of them, sorry. I'm a hypocrite because whilst I love animals I'm a sucker for a chicken burger.
-
I wasn't having a go at you Freddy, just saying my bit.
One day maybe you will get the chance to try some of Mrs Data's gorgeous food and you might think again ;D
Peace ;)
-
I knew you weren't having a go my friend. :)
-
If only all animals were free it would be a lot easier. We could classify our meals as a Mother Nature's will, hoping that everything is fine. At least our kids will have the fancy future, I'm sure of it. I've heard of Mc Donalds programs of having natural environments with hedge mazes and everything for their cows yet now.
:valentine2:
-
There are more humane ways to kill animals. I recently saw a news story on processing plants for chickens and i could not believe they were still killing them that way. When I was young i worked at one and it was disgusting the poor chickens were not killed first they just went right into the plucking machine still alive. I did not stay working there long. My mothers family was farmers my mom was not of course being married to a chemist but they were so i was familiar with how animals were killed. Their methods were humane it was over quickly so quick that they did not even have time to make noise about it.
We don't let murders suffer so we at least afford our food the same respect.
-
I don’t normally comment on subjects like this, but I thought I’d give my bobs worth. Firstly let me say that I do care about animals… but a little perspective is needed.
All animals (humans included) are the sum of their experiences. Colin Blakemore proved that even the vision system arises from the stimulus animals receive after birth.
WARNING: Don’t watch if you’re the sensitive type… https://www.youtube.com/watch?v=QzkMo45pcUo (https://www.youtube.com/watch?v=QzkMo45pcUo)
Basically cats which are reared in an environment with only horizontal lines to view for the first three months can only ever see horizontal lines.
All the cortices work like this. If an animal has never known or experienced freedom etc, then how can it miss it? We must be careful not to anthropomorphise and impose our own understanding of pain and suffering on the animals we cull.
It’s probably worse for a free range animal that has experienced sunshine, freedom and the companionship of its fellow species than it is for a battery hen when it’s time comes.
The human race would not be where it is today if we didn't eat meat. Vegetarianism is a modern indulgence people can pursue because we have the knowledge required to supplement our diets. If you had tried it even a few hundred years ago… you would have died.
-
I didn't watch that video... :(
Just a thought for all of you: there is a difference between a kill with a smile on your face or a murder with a tear in your eye.
-
I didn't expect the link to show the first frame from the video… soz for that.
-
I didnt watch it either, maybe me and Ivan have something in common.
There was this nasty video of this guy cyborgizing a real cockroach, its horrible, I dont think ai really needs it, I wouldnt call it easy, but I dont think operating on live animals brains helps anything, I think its just logic your born with and some hardware/software skill is all someone would need.
-
This video is very interesting, as in - gave interesting results. I've always thought that we are not born with the ability to recognize some patterns and I guess we probably wouldn't know how to react to a say, ball propelled towards us - until we witness what it can do ( damage to another object, creature...).
-
I watched it. I was glad the kitten had a normal life after acquainting itself with reality. Putting it in that box with the lines made me sad though I have to admit. There have been far worse things done to animals in the name of science and curiosity though.
-
ball propelled towards us - until we witness what it can do
Oh… this opens the whole can of worms… about free will.
I’m going to cause an argument by saying… the first time we experience a situation, we can’t exercise free will. Also, I don’t think there is such a thing as memories… only correct predictions.
Hear me out…
We don’t actually live in the moment; if we did we would be constantly lagging behind reality. Our brain forms predictions using our internal virtual model of the world as we think it works.
If we are forming predictions they must be based on previous knowledge and experience. You must have the experience/ patterns in place to recognize the current experience in the first place.
So if a situation arises that we have no previous experience of we can’t form a prediction/ choice of predictions based on free will… we have no choices to choose from.
If you are in a bank and someone shouts… FREEZE… you do. The only way to override this is through training and experiencing all the possible combinations likely in life. Hence army and police training! Also causes rabbits and deer to stop in your headlights etc… no previous experience to form a prediction/ reaction too.
Therefore if memories do exist they do so only as proven predictions, and there is no such thing as free will, in the first instance of an experience.
-
Yup, just watch babies. They don't react to much at first. They just sit or lay there wondering about things blissfully unaware of the dangers in the world. Which is why we have to look after them.
That moment of not knowing what to do in some new situation must be familiar to everyone. I'd say we are not entirely helpless though, but I agree that past experience is probably the key to working out what to do next. Comparison and assessment before taking action - which could take time.
What about instinct though ?
-
It does seem that instinct mostly applies to animals more than people.
A Salmon knows to spawn upstream, bees know how to work and live in a colony although they are highly intelligent in the insect world, as are ants.
Why do some birds fly south for the winter and how do they know when to return in the spring?
Why does a squirrel bury nuts (although it can only remember where a small percentage are hidden)?
Some people might say they distrust an individual because there's just "something" about him / her that they can't put their finger on...instinct.
Are humans or animals perhaps reading or interpreting other signs or signals then using an internal prediction of outcome within our brains to form a thought called an instinct?
An instinct is a behavior or feeling that is acted upon and is inborn in an animal or maybe a person, although I still contend that animal instinct far outweighs human instinct.
-
An instinct is a predisposition to a particular reflex action or trait that has been passed down in the connectome wiring/ DNA of an animal that serves survival of the fittest.
All the humans that don’t raise their heads when they feel water on their face… are dead. lol
This of course is the theory behind use of Genetic Algorithms, were random mutation is tested for fitness and the best are combined.
-
Babies have a sucking instinct. They also will grab on to something, like a human finger. This keeps them from falling off their mother. I think jumping and screaming after being startled might be an instinct too. Some people will suddenly even hit the person who scared them. The sex drive and the sleep cycle might also be instinctual, in a way. It is hardwired information injected deep within our brain and body. Perhaps breathing too. I guess it depends on how broad your definition of instinct is.
Although, I don't ever recall ever going on a migration, or hibernating. I think getting a bad feeling about a person is closer to intuition. Getting a feel for someone is suppose to be a function of your right hemisphere.
-
I guess that works if we define instinct as knowledge or skills we are born with. Instinct is actually information essential (to a degree) for our survival which has been passed to us from past generations, and also refined as it has passed from one predecessor to the next.
It is also pretty interesting too - it helps babies not injure themselves when they fall (mostly while learning to walk) - they instinctively fall on their butts. It is a concept we have been taught at Bujinkan (martial art organization) - we are always told to first lower ourselves / bend, and before any fancy roll or regular fall the first step is to bend with a tendency to sit down.
Think I kind of messed up the words but I believe you will get my point.