I, Android
Aug. 4th, 2004 12:35 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
strredwolfI saw I, Robot yesterday... and it made me wonder what would happen if we didn't use the Three Laws, which are very very subjective.
In a few universities, they're "growing" an AI. They're treating an AI like a baby and building up a set of basic assumptions, making it think human... and making sure it thinks like we think so if the Three Laws were put in, it would work right... but we're not going to put them in, aren't we?
Let's put this all into the Canmephian universe. Before RedWolf came back, there were Mk I AI's (Insect-like minds). Simple stuff. Afterwards, RedWolf realized shi was an organic AI, a virtual Mk V. So shi hooked up with some researchers and aided in creating the Mk II class, which are basic bit-bucket computers like what you're sitting infront of, controlling a shell of a body. RedWolf created, and maintains, the Basic Assumption List, which is in all AI's from Mk II on, including the Mk. III that is Zero (Xili's mecha) and the Mk IV that is FireTiger. Mk III's would be similar to the NS-5, but not Three Laws Safe. There is saftey in the Basic Assumption List, though, because it is known what is it (it's an open source list).
Some may say OS upgrades are an epiffany.
However, there is a back door. Infact, two. RedWolf not only maintains the BAL, but also is root/admin of the AI's. It's the weak link in the chain. What fun.
In a few universities, they're "growing" an AI. They're treating an AI like a baby and building up a set of basic assumptions, making it think human... and making sure it thinks like we think so if the Three Laws were put in, it would work right... but we're not going to put them in, aren't we?
Let's put this all into the Canmephian universe. Before RedWolf came back, there were Mk I AI's (Insect-like minds). Simple stuff. Afterwards, RedWolf realized shi was an organic AI, a virtual Mk V. So shi hooked up with some researchers and aided in creating the Mk II class, which are basic bit-bucket computers like what you're sitting infront of, controlling a shell of a body. RedWolf created, and maintains, the Basic Assumption List, which is in all AI's from Mk II on, including the Mk. III that is Zero (Xili's mecha) and the Mk IV that is FireTiger. Mk III's would be similar to the NS-5, but not Three Laws Safe. There is saftey in the Basic Assumption List, though, because it is known what is it (it's an open source list).
Some may say OS upgrades are an epiffany.
However, there is a back door. Infact, two. RedWolf not only maintains the BAL, but also is root/admin of the AI's. It's the weak link in the chain. What fun.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2004-08-04 09:58 am (UTC)I would have more to say, but I didn't even know that Redwolf's brain was an AI shell running a bio emulation (if that's how you would put it). I'm still a bit confuzzled about those origins, but haven't had the time to look into it, or at least, haven't been clicking the right links.
I've only read about two projects actually "growing" AI. The first is more of an evolutionary iteration growth, where there are satellite robots that are made to perform a task, like walking on a floor. But, if it reaches a stair, it can't go any further, so the mother robot builds another that will go up a stair, and walk on the floor, until it reaches another obstical, such as a wall, in which case the mother has to build one that will walk, go up stairs, and turn at a wall.... and so on and so forth. It's more of an evolution of kinesthetic responses, more than anything.
I heard rumors of another type that was a bit less physical and more computer based, but I can't remember if those are true or not.
no subject
Date: 2004-08-04 10:01 am (UTC)I personally know that the term "Robot" is derived from a word that means "Slave". Not sure how automaton would apply as such. It's an odd concept all around.
But how did the word Android evolve? And why do you use that term instead of others?
I need to google the history of that word when I have the time.
no subject
Date: 2004-08-04 10:23 am (UTC)no subject
Date: 2004-08-04 10:35 am (UTC)Err.. Okay, now my brain has seized on that last statement, ah well :)
no subject
Date: 2004-08-04 03:49 pm (UTC)If not, then i guess i'll miss out the year of the Gundam mecha. hahahahahahah :))
Read the book
Date: 2004-08-13 03:42 pm (UTC)Quite a few storys of robots malfuctioning.
Egnr of Neko Homeworld could never have the three laws.
Energy matrix are self evolving and eventually evolve past the original programming so any attempt to include safeguards would be futile as the matrix (energy matrix version of a binary) would be recycled for more "useful" functions.
There is a basic safeguard. Once the matix start acting outside set peramiters an emergency eject procedure is used..
(Depends on who wrote the matrix as to what that procedure is)
Most just drop em into space where they burn out and die...
Kinda sad really...