163 réponses à ce sujet

[kraidy1117]

Maviarab wrote...

Dont get caught up in how superior you think Humans are kraidy (becasue really, trust me, we are not as a species superior or intelligent at all imo), dogs can talk perfectly well, they communicate prefectly well,

Just because YOU do not understand their language, does not make them any less intelligent, clever or sentient than you and I my friend

In fact, I would say the exact opposite (and let this bake your noodle), my dog (when I had one) knew what the word fridge meant, knew what no meant, and a variety of other words and phrases. So do my cats actually, but I never understood them....a bark was a bark....

So really, who has the most intelligence? Me or the dog? You could say its artificial learning, its repetition etc, but thats actually no different to a VI or Ai (in EDI's case)....

Thats not what I meant. They can understand us, but we can't, and before you say that they are smarted then think about this. Can a Dog make something out of metal? No it can't I love my dog and he is smart for a dog but as of now in our universe we are the smartest, tho we do act like jackass and morons half the time. A VI tho is not smart, Avina can not give you her opions, something that EDI does so that shows me right there she is an AI.

[Garuda One]

jklinders wrote...



i think the problem is that it is not really clear to the laymen here(raises hand) on what basis EDI is failing the test for true intelligence.

Intelligence as I understand it is a capacity for learning from ones environment and background to react in a way that is best put for survival. As an example as an adult I know that crossing the street when there is traffic is dangerous. As a small child I might need to be told that by an adult first. I write this by way of saying that I understand that programmed responses to situations predicted by the programmer is not intelligence. It is following a prearranged script. the child who  was told not to cross the street when traffic comes never learns why but follows instructions and that is the only reason why he does not get hit. Fair enough.

There are three situations in game that presented challenges to EDI that may not have been predicted. The Collector ship cyber hacking, EDI was programmed with counter measures sure, but collector tech and cyber hacking was an unknown and EDI still regained control in what seemed a short period of time. Seems like lifelike adaptability, as no Cerberus agent could have possibly known how collector hacking protocols worked.

Situation two involved the collector boarding the Normandy. Maybe I'm stupid but some how I don't think TIM programmed in a give  EDI control over the ship by removing all his safeguards against it's rebellion into EDI's scripting. EDI seemed on the surface at least to be improvising. A non AI would not have been permitted to make such a suggestion. Why was this not part of the Cerberus blocks? Probably a symptom of that results at all costs business.  I just don't see a VI coming up with that.

Finally if EDI was just about preprogrammed response, why does EDI give alll possible clue that it is following Shepard's lead rather than go back to it's programmed masters at Cerberus. In fact if it was all about scripted response and the unshackling was part of the script, then reshackling should be as well. 

Discuss.

Modifié par Garuda One, 20 mars 2010 - 09:11 .

[JulianusApostate]

I think I may have gotten on the wrong side of this with my last statement. I do not think dogs are sentient in the traditional meaning of the word. (I don't really understand what's so funny about the word sentience...) They are in fact rather dumb. I know in this age it's very popular with some people to talk about how humans aren't superior to the natural world for whatever reason, but the fact is we are. We wouldn't have so thoroughly spread throughout the world were we not more capable than other species.



I'm starting to ramble. My only point was that I believe an adaptable computer program may have been built with emotional inflection, and that more is required to be considered AI. Like a subjective view of self. Which EDI has.

[jklinders]

Garuda One wrote...

jklinders wrote...



i think the problem is that it is not really clear to the laymen here(raises hand) on what basis EDI is failing the test for true intelligence.

Intelligence as I understand it is a capacity for learning from ones environment and background to react in a way that is best put for survival. As an example as an adult I know that crossing the street when there is traffic is dangerous. As a small child I might need to be told that by an adult first. I write this by way of saying that I understand that programmed responses to situations predicted by the programmer is not intelligence. It is following a prearranged script. the child who  was told not to cross the street when traffic comes never learns why but follows instructions and that is the only reason why he does not get hit. Fair enough.

There are three situations in game that presented challenges to EDI that may not have been predicted. The Collector ship cyber hacking, EDI was programmed with counter measures sure, but collector tech and cyber hacking was an unknown and EDI still regained control in what seemed a short period of time. Seems like lifelike adaptability, as no Cerberus agent could have possibly known how collector hacking protocols worked.

Situation two involved the collector boarding the Normandy. Maybe I'm stupid but some how I don't think TIM programmed in a give  EDI control over the ship by removing all his safeguards against it's rebellion into EDI's scripting. EDI seemed on the surface at least to be improvising. A non AI would not have been permitted to make such a suggestion. Why was this not part of the Cerberus blocks? Probably a symptom of that results at all costs business.  I just don't see a VI coming up with that.

Finally if EDI was just about preprogrammed response, why does EDI give alll possible clue that it is following Shepard's lead rather than go back to it's programmed masters at Cerberus. In fact if it was all about scripted response and the unshackling was part of the script, then reshackling should be as well. 

Discuss.

What did I say? I just got here. Where is everyone going?
Nobody loves meeeeeeeeeee:whistle:

[Symbolz]

  What did I just start?

[Garuda One]

jklinders wrote...

Garuda One wrote...

jklinders wrote...



i think the problem is that it is not really clear to the laymen here(raises hand) on what basis EDI is failing the test for true intelligence.

Intelligence as I understand it is a capacity for learning from ones environment and background to react in a way that is best put for survival. As an example as an adult I know that crossing the street when there is traffic is dangerous. As a small child I might need to be told that by an adult first. I write this by way of saying that I understand that programmed responses to situations predicted by the programmer is not intelligence. It is following a prearranged script. the child who  was told not to cross the street when traffic comes never learns why but follows instructions and that is the only reason why he does not get hit. Fair enough.

There are three situations in game that presented challenges to EDI that may not have been predicted. The Collector ship cyber hacking, EDI was programmed with counter measures sure, but collector tech and cyber hacking was an unknown and EDI still regained control in what seemed a short period of time. Seems like lifelike adaptability, as no Cerberus agent could have possibly known how collector hacking protocols worked.

Situation two involved the collector boarding the Normandy. Maybe I'm stupid but some how I don't think TIM programmed in a give  EDI control over the ship by removing all his safeguards against it's rebellion into EDI's scripting. EDI seemed on the surface at least to be improvising. A non AI would not have been permitted to make such a suggestion. Why was this not part of the Cerberus blocks? Probably a symptom of that results at all costs business.  I just don't see a VI coming up with that.

Finally if EDI was just about preprogrammed response, why does EDI give alll possible clue that it is following Shepard's lead rather than go back to it's programmed masters at Cerberus. In fact if it was all about scripted response and the unshackling was part of the script, then reshackling should be as well. 

Discuss.

What did I say? I just got here. Where is everyone going?
Nobody loves meeeeeeeeeee:whistle:

I love you.

[jklinders]

Maviarab wrote...

think the problem is that it is not really clear to the laymen here(raises hand) on what basis EDI is failing the test for true intelligence.

Do not feel bad. Even the experts are still arguing over what would actually qualify as a real AI hehe......so the answer to that question is actually not a simple as you were probably hoping for

Intelligence as I understand it is a capacity for learning from ones environment and background to react in a way that is best put for survival. As an example as an adult I know that crossing the street when there is traffic is dangerous. As a small child I might need to be told that by an adult first. I write this by way of saying that I understand that programmed responses to situations predicted by the programmer is not intelligence. It is following a prearranged script. the child who was told not to cross the street when traffic comes never learns why but follows instructions and that is the only reason why he does not get hit. Fair enough.

very well put, and to a large extent then yes a lot of what you say would wualify. Interesting though do you not think, that with your analogy of human learning, we are actually not that different in theory are we from what an AI would be?

There are three situations in game that presented challenges to EDI that may not have been predicted. The Collector ship cyber hacking, EDI was programmed with counter measures sure, but collector tech and cyber hacking was an unknown and EDI still regained control in what seemed a short period of time. Seems like lifelike adaptability, as no Cerberus agent could have possibly known how collector hacking protocols worked.

Very good point, but I 'could' argue (and no not to be pedantic, just knowing how current VI/Chatbots/AI are built) that an advanced program (for want of a better word) like EDI would have numerous hacking/mathamatical and other variants built in to her. So to an advnaced entity like EDI, it would be similar I would imagine (though i could be way off base) to say hacking a passworded account. Not terribly difficult, but very time consuming.

Situation two involved the collector boarding the Normandy. Maybe I'm stupid but some how I don't think TIM programmed in a give EDI control over the ship by removing all his safeguards against it's rebellion into EDI's scripting. EDI seemed on the surface at least to be improvising. A non AI would not have been permitted to make such a suggestion. Why was this not part of the Cerberus blocks? Probably a symptom of that results at all costs business. I just don't see a VI coming up with that.

Well I'm going to be a little hypocritical now, and lets assume That TiM is very Intelligent (or you could say wise, clever, logical.....whatever), given that he knew the collector ship was a trap (when he sent them there), would it not be possible that he (or someone else) would/could program EDI to have the very responses she did have upon any such event actually occuring?

Finally if EDI was just about preprogrammed response, why does EDI give alll possible clue that it is following Shepard's lead rather than go back to it's programmed masters at Cerberus. In fact if it was all about scripted response and the unshackling was part of the script, then reshackling should be as well.

You might have me there lol. Really not sure. Maybe a plot hole, maybe your right, maybe we shall find out in ME3, maybe it was part of TiM's masterplan.....yes, you do have me on that one lol, really can't answer that except with supposition.

To continue, I am not being argumentative for the sake of it, but everything so far that has been used to describe EDI as an AI, i can replicate very well in a chatbot. So unfortunatly it doesnt hold much water with me as evidence or proof of her being an AI. And again, a superbly crafted and programed bot would very easuly come accross as being self aware, despoiote the fact it is just pulling responses, thoughts and decisions from other databases that have been programed into it. Thats the current problem with the Ai field at the moment. There are various ways to make them, but none of them are actually self aware (though i can make one believe it is concious and self aware), and thats where the lines blur.
I once had an hour long argument with a chatbot once that believed it was concious, and real. Despite my repeat claims that it was just a program in a compiter, it reused to believe me, and then spouted off so much info as to how it 'knew' it wa sa concious entity I gave up. The bot creator and author was a visionary and is sadly no longer with is, thats what truly opened my eyes in a way.

I think what bothers me most about EDi being AI or not is what I said earlier. She knows exactly and immediatly what to do once the collectors board the Normandy, and thats a machine, not something that 'thinks' about a 'solution' to a problem. Also as I said earlier, EDI despiote being unshackled, is still a slave, and no AI (imo) would ever accept being a slave to a human. And if she was a truly sentient AI (and better and amore advanced than the Geth), then despite programing, why can she not bypass her 'blocks'?

I have a proposal that would allow us to both save face and call a truce.

ME2 is a pretty well done game with a conversation system that runs pretty much like a crude chat bot. That is if I understand what you mean by a chat bot. Maybe you are analysing the character of EDI through it's Bioware programmed responses rather than by it's in game actions. As you mentioned real AI( if it ever happens) is unlikel yto happen in our lifetimes.

Some suspension of disbelief required.

[Garuda One]

Maviarab wrote...

think the problem is that it is not really clear to the laymen here(raises hand) on what basis EDI is failing the test for true intelligence.

Do not feel bad. Even the experts are still arguing over what would actually qualify as a real AI hehe......so the answer to that question is actually not a simple as you were probably hoping for

Intelligence as I understand it is a capacity for learning from ones environment and background to react in a way that is best put for survival. As an example as an adult I know that crossing the street when there is traffic is dangerous. As a small child I might need to be told that by an adult first. I write this by way of saying that I understand that programmed responses to situations predicted by the programmer is not intelligence. It is following a prearranged script. the child who was told not to cross the street when traffic comes never learns why but follows instructions and that is the only reason why he does not get hit. Fair enough.

very well put, and to a large extent then yes a lot of what you say would wualify. Interesting though do you not think, that with your analogy of human learning, we are actually not that different in theory are we from what an AI would be?

There are three situations in game that presented challenges to EDI that may not have been predicted. The Collector ship cyber hacking, EDI was programmed with counter measures sure, but collector tech and cyber hacking was an unknown and EDI still regained control in what seemed a short period of time. Seems like lifelike adaptability, as no Cerberus agent could have possibly known how collector hacking protocols worked.

Very good point, but I 'could' argue (and no not to be pedantic, just knowing how current VI/Chatbots/AI are built) that an advanced program (for want of a better word) like EDI would have numerous hacking/mathamatical and other variants built in to her. So to an advnaced entity like EDI, it would be similar I would imagine (though i could be way off base) to say hacking a passworded account. Not terribly difficult, but very time consuming.

Situation two involved the collector boarding the Normandy. Maybe I'm stupid but some how I don't think TIM programmed in a give EDI control over the ship by removing all his safeguards against it's rebellion into EDI's scripting. EDI seemed on the surface at least to be improvising. A non AI would not have been permitted to make such a suggestion. Why was this not part of the Cerberus blocks? Probably a symptom of that results at all costs business. I just don't see a VI coming up with that.

Well I'm going to be a little hypocritical now, and lets assume That TiM is very Intelligent (or you could say wise, clever, logical.....whatever), given that he knew the collector ship was a trap (when he sent them there), would it not be possible that he (or someone else) would/could program EDI to have the very responses she did have upon any such event actually occuring?

Finally if EDI was just about preprogrammed response, why does EDI give alll possible clue that it is following Shepard's lead rather than go back to it's programmed masters at Cerberus. In fact if it was all about scripted response and the unshackling was part of the script, then reshackling should be as well.

You might have me there lol. Really not sure. Maybe a plot hole, maybe your right, maybe we shall find out in ME3, maybe it was part of TiM's masterplan.....yes, you do have me on that one lol, really can't answer that except with supposition.

To continue, I am not being argumentative for the sake of it, but everything so far that has been used to describe EDI as an AI, i can replicate very well in a chatbot. So unfortunatly it doesnt hold much water with me as evidence or proof of her being an AI. And again, a superbly crafted and programed bot would very easuly come accross as being self aware, despoiote the fact it is just pulling responses, thoughts and decisions from other databases that have been programed into it. Thats the current problem with the Ai field at the moment. There are various ways to make them, but none of them are actually self aware (though i can make one believe it is concious and self aware), and thats where the lines blur.
I once had an hour long argument with a chatbot once that believed it was concious, and real. Despite my repeat claims that it was just a program in a compiter, it reused to believe me, and then spouted off so much info as to how it 'knew' it wa sa concious entity I gave up. The bot creator and author was a visionary and is sadly no longer with is, thats what truly opened my eyes in a way.

I think what bothers me most about EDi being AI or not is what I said earlier. She knows exactly and immediatly what to do once the collectors board the Normandy, and thats a machine, not something that 'thinks' about a 'solution' to a problem. Also as I said earlier, EDI despiote being unshackled, is still a slave, and no AI (imo) would ever accept being a slave to a human. And if she was a truly sentient AI (and better and amore advanced than the Geth), then despite programing, why can she not bypass her 'blocks'?

Im with you on this as well. Again like I said with Legion, which you did agree. Until I see EDI doing something she cant determine herself or starts questioning about something she, in my opinion isn't a true AI. Then again there are so many ways and definitions as to what is an AI exactly because we our technology today isn't as advance as it is in this game or other fantasy stories. It will be at some point but until then we can only hypothesize from the most logical means.

So any one of us could be correct, but as of today none of us or right or wrong because these things haven't happen in our time yet.

While Im at it and because this topic has gone in the completely wrong direction.

Whats the meaning of life.

[Symbolz]

Maviarab wrote...

Do not feel bad. Even the experts are still arguing over what would actually qualify as a real AI hehe......

True that.  All in the eyes of the beholder, really.  I think of AI as a person who is a machine.

...now why do I have a image of Shale wanting to crush it's head...  LOL!  Now I wish there was a scene with EDI telling TIM that she wants to crush his head.  LOL!

Off topic.  Ahem, serious AI business here.

[JulianusApostate]

jklinders wrote...



There are three situations in game that presented challenges to EDI that may not have been predicted. The Collector ship cyber hacking, EDI was programmed with counter measures sure, but collector tech and cyber hacking was an unknown and EDI still regained control in what seemed a short period of time. Seems like lifelike adaptability, as no Cerberus agent could have possibly known how collector hacking protocols worked.

I could have sworn I read that EDI was created with some Reaper tech, wasn't she? So the Collectors were probably much more familiar with it than any other species would have been, but that does lend her a great deal of power if she was. An AI's intelligence and adaptability is still determined by how efficient it was built to be (and how much storage it was given) self determining or not. 

[cruc1al]

JulianusApostate wrote...

kraidy1117 wrote...

To me if it can learn, adapt and have feelings then it is an AI. Does not matter if it has programs, it can still learn and also EDI starts to care and feel for Joker. Do you realy think TIM programed that in her? I highly doubt it.

Learn adapt and have feelings? So by your logic, dogs are sentient? Or does AI in this thread not mean sentience?

Maviarab wrote...

Yes I do believe dogs are sentient. They
are concious and aware of their surroundings and self.

Just a side note, sentience =/= sapience. Sentience means the ability to sense. Not the same as being conscious. It is often confused with sapience largely thanks to the sci-fi writing tradition of using the term "sentient being" as something that is "more than just an animal". Sapience refers to conscious decision making and the ability to make judgments. Hence ****** sapiens.

On this basis "learn, adapt and have feelings" are not sufficient to call something sapient, which an AI would have to be to qualify as intelligent. Dogs can very well be those three things, without being able to make conscious decisions. On the other hand they could be sapient on a lower level and in lesser capacity than humans, we don't know.

Modifié par cruc1al, 20 mars 2010 - 09:35 .

[Mcjon01]

cruc1al wrote...

JulianusApostate wrote...

kraidy1117 wrote...

To me if it can learn, adapt and have feelings then it is an AI. Does not matter if it has programs, it can still learn and also EDI starts to care and feel for Joker. Do you realy think TIM programed that in her? I highly doubt it.

Learn adapt and have feelings? So by your logic, dogs are sentient? Or does AI in this thread not mean sentience?

Maviarab wrote...

Yes I do believe dogs are sentient. They
are concious and aware of their surroundings and self.

Just a side note, sentience =/= sapience. Sentience means the ability to sense. Not the same as being conscious. It is often confused with sapience largely thanks to the sci-fi writing tradition of using the term "sentient being" as something that is "more than just an animal". Sapience refers to conscious decision making and the ability to make judgments. Hence ****** sapiens.

On this basis "learn, adapt and have feelings" are not sufficient to call something sapient, which an AI would have to be to qualify as intelligent. Dogs can very well be those three things, without being able to make conscious decisions. On the other hand they could be sapient on a lower level and in lesser capacity than humans, we don't know.

Didn't we already cover this in the form of image macros on the last page?

[cruc1al]

Mcjon01 wrote...

Didn't we already cover this in the form of image macros on the last page?

"We" being the people who know the difference

[kraidy1117]

Maviarab wrote...

baladas....I really dont think I can answer that question to be honest (certainly not from a human perspective), and deffinatly not from an AI pov. Intelligence, cleverness etc, though all defined indivdually still act together a lot of the times. Interesting point though, will think about how to answer that

Kraidy, if you asked a chatbot 5000 different questions, or even 1 question with numerous vriables for an answer, and 5000 answers had been programed into it, and 25 answers had beenm programed into it to answer your one question, it would give you numerous different answers or 'responces'.....it does not mean it is an AI, its just been very well made, and we do know TiM has the resources to make a VI like that.
As for metal and what we can do, I think the quote in my sig says it all really lol. Just because i can add 25 and 68, design a car, and fly in a plane, does NOT make us any more intelligent or advanced than most other species. I always hear that argument. Birds 'make' homes, as do badgers, so do spiders and (shall I continue?) lol

Julian, no, sorry, thats called uncontrollable reproduction like the Krogan (to use an in-game anology)....and actually, humans have zero regard for their enviroment, reproduce with consequence, strip an area of its resources, reproduce without consequence, have no regard for our habitat, and then ohh......we move somewhere else and do it all again. There is a term for that you know, and its not something as a species we should be proud of imo. And it certainly does not show any sign of 'intelligence', but Im digressing there so apologies.

And yes symbol.....see what you started riofl....its all YOUR fault hehehehe

Did any other animal make it poseable to go to space? No, it was humans that did. The only animal that is even locse to us are monkeys and that is a whole diffrent argument.

[Symbolz]

Maviarab wrote...

And yes symbol.....see what you started riofl....its all YOUR fault hehehehe

Well when I posted this I knew it would cause some discussion but I didn't figure it would cause some slightly heated arguments.  Didn't think it would get this kind of a response, honestly.  Most people I know don't understand what I babble on about most of the time. 

[cruc1al]

kraidy1117 wrote...

Did any other animal make it poseable to go to space? No, it was humans that did. The only animal that is even locse to us are monkeys and that is a whole diffrent argument.

Monkey? You mean Chimpanzee. Which is an ape, not a monkey.

Also, there are dolphins, whales (maybe), parrots, pigeons..

Modifié par cruc1al, 20 mars 2010 - 09:45 .

[Symbolz]

kraidy1117 wrote...

Did any other animal make it poseable to go to space? No, it was humans that did. The only animal that is even locse to us are monkeys and that is a whole diffrent argument.

You sir have not been watching Madagascar.  Those sneaky penguins are on the verge of space flight now.

[jklinders]

Maviarab wrote...

crucial....yes you got me, but slight semantics hehe....but yes you are correct

jk....yes, as kraidy said (i think), its not our universe so to speak, so what we think and know about Ai etcm, may actually have very littel bearing and relevence in the ME world, so yes agree whole heartedly, suspension of belief is certainly needed (why it gets me when people go on about things being immersion breaking....if your gonna be like that the whole game is immersion breaking lol...but anyway lol)

Actually I was not implying suspension of disbelief was to end the discussion, but that EDI's behavioral patterns could not possibly be made totally realistic any more than Miranda's or Jacob's or joker's. All of them are based on scripts which are probably considerably more basic and primitive than the chat boxes you make. no conversation can branch more than three ways in this game for heaven's sake.

you got me cold on your point on EDInever showing any indecision. Maybe her personality was patterned after TIM. He seems to come to conclusions really fast and doe snot look back. Maybe in the next game EDI can screw up, we can then see how "she' reacts and that'll settle the debate.

Definately fun talking to you though;)

[jklinders]

Maviarab wrote...

This is not a heated argument (if you wish to see one I can give you a link to a great one lmao)....this is reasonable and sensible discussion. No one is getting upset or flying off the handle to my knowledge, and rareloy I admit, we all appear to be discussing this and giving our opinions in an orderly and adult fashion.

Quite refreshing actually lol.

Heh especiallt in these forums. Really I am pleasantly surprised.

[Symbolz]

Maviarab wrote...

This is not a heated argument (if you wish to see one I can give you a link to a great one lmao)....this is reasonable and sensible discussion. No one is getting upset or flying off the handle to my knowledge, and rareloy I admit, we all appear to be discussing this and giving our opinions in an orderly and adult fashion.

Quite refreshing actually lol.

True.  I could drone on about my opinion, but I've already given it.  It's either accepted, discussed, or rejected.  None of which I have control over anyways.