Most Convincing AIs in Literature?

[Dimwit]

If you are going to do some quoting then I have a small correction, in Ringo's Aldenata Sagas the PDA's where called Buckleys after the main overlay. These were local (Terran) manufacture and were less capable than the true alien AI PDA's called Aides.

Another one might be McCaffrey's the Ship Who Sang series. While not "true" AI's since they were derived from people, McCaffrey calls them AI's.

[fjtorres]

The brainships are actually people; humans equipped with prosthesis that happen to be spaceships. Calling them AIs is stretching the concept past the breaking point.

[charleski]

Iain M. Banks

Like most AIs in fiction, his creations are often idealised forms of humanity. But not always:

'Never forget I am not this silver body, Mahrai. I am not an animal brain, I am not even some attempt to produce an AI through software running on a computer. I am a Culture Mind. We are close to gods, and on the far side.
'We are quicker; we live faster and more completely than you do, with so many more senses, such a greater store of memories and at such a fine level of detail. We die more slowly, and we die more completely, too. Never forget I have had the chance to compare and contrast the ways of dying.'
...
There was an utter coldness and a ferocious, infinite intelligence in those dark eyes. Something without mercy or illusion was staring straight into his soul and finding it not so much wanting, as absent.

[Look to Windward]

[rkomar]

It's not recent, but John T. Sladek's "The Reproductive System" (aka "Mechasm" in North America) is a book about the evolution of simple, autonomous robots into a larger, intelligent system. It's sci-fi humour from the 60's, but worth a read today. I doubt if it's available as an ebook, though, as even paper copies of Sladek's books are getting mighty thin on the shelves.

[Solitaire1]

Originally Posted by fjtorres

Of course, quoting Chobits opens up the door for John Smith, the Red Tornado, and his siblings. Also Brainiac, Kilg%re, Ultron, Jocasta, Vision, Cliff Steele, the Metal Men, and a zillion killer AI's robots from both DC and Marvel.

The Metal Men alone have many enemies who could fall under the heading of AI, including:

- The Gas Gang: A robot team made of gas, rather than metal.

- The Inheritor: A plastic robot with all of the Metal Men's abilities and none of their weaknesses.

- Chemo: A man-shaped vat of toxic waste that gained pseudo-life with a very limited intelligence.

On the subject of novels featuring an AI, one to mention is "Virtual Girl" by Amy Thomson. A man builds a female robot as a companion and it develops a true AI.

Since "Chobits" has been mentioned, I'll mention a few other anime that features at least one AI or possible AI: A.D. Police, All-Purpose Cultural Cat-Girl Nuku-Nuku, Armitage III, Astro Boy, Big O, Bubblegum Crisis, Dragonball Z, Excel Saga, Ghost In The Shell, Macross Plus, Mahoromatic, Saber Marionette J, and Steel Angel Karumi.

I must also mention "Angelic Layer" (AL). Although the anime of "Chobits" doesn't mention it, both AL and "Chobits" are set in the same universe and the events of AL (such as the development of the angels) lead to the creation of the persocoms.

[Nancy Fulda]

Okay, the article's chugging along. I think I can fit Mycroft and Commander Data in there.

Not sure whether I'll find a spot to sneak in Do Androids Dream of Electric Sheep or any of the other excellent suggestions in this thread. I do have to leave space to discuss the mechanics of neural networks and Bayesian learning, after all...

[Nancy Fulda]

Originally Posted by fjtorres

The brainships are actually people; humans equipped with prosthesis that happen to be spaceships. Calling them AIs is stretching the concept past the breaking point.

Agreed. Although I do love the series.

[alexandree.spier]

Hi Nancy,

I thought you might be interested in Benedict Anderson's "Imagined Communities". It's an old book but revised many times... I was just wondering.

[Nancy Fulda]

I'll take a look at it. Thanks!

Edited to Add: Um. Is there more than one Benedict Anderson with a book by that title? Because the one I found is about nationalism and does not appear to have anything to do with AI at all.

[Penforhire]

There is the Eschaton of Charlie Stross' "Singularity Sky" & "Iron Sunrise." Not a ton of detail there and I don't recall if Stross was even explicit about the big E being artificial. But I liked the idea that a future super-intelligence might gain some sort of vision-of-the-past ability and become very possessive of, or guarded about, use of technologies that might inhibit its own future creation.

[Aidan]

Another interesting one no-one's mentioned is the AI in Mieville's Perdido Street Station. It wasn't a primary part of the plot but I thought it was a great exposition of how an AI could come into being and develop.

[starrigger]

Speaking from a writer's point of view, it's almost impossible to write future-oriented SF without involving AIs. The interesting question is how do you use them. I've had a lot of fun with the development of some of my own AIs, including the robots Napoleon and Copernicus in the Chaos books, who have gradually become sentient through alien-enhanced upgrades. After a while, they've become characters who grow throughout the stories just as the other characters do.

Those robots were not created--by their fictional creators--specifically to have personalities; the personalities just evolved with their experience. An AI I wrote about in the 1980s, in a novel called The Infinity Link, was designed--by its fictional creators--to be an artificial human personality. That led to totally different sorts of consequences in its interactions with human characters, including one whose mind and personality had been uploaded.

Which raises the question, if a real human intelligence and personality is uploaded into a computer matrix, is it then an AI? Seems to me you could argue either way.

[EatingPie]

Originally Posted by starrigger

Speaking from a writer's point of view, it's almost impossible to write future-oriented SF without involving AIs. The interesting question is how do you use them. I've had a lot of fun with the development of some of my own AIs, including the robots Napoleon and Copernicus in the Chaos books, who have gradually become sentient through alien-enhanced upgrades. After a while, they've become characters who grow throughout the stories just as the other characters do.

Those robots were not created--by their fictional creators--specifically to have personalities; the personalities just evolved with their experience. An AI I wrote about in the 1980s, in a novel called The Infinity Link, was designed--by its fictional creators--to be an artificial human personality. That led to totally different sorts of consequences in its interactions with human characters, including one whose mind and personality had been uploaded.

Which raises the question, if a real human intelligence and personality is uploaded into a computer matrix, is it then an AI? Seems to me you could argue either way.

I'll bite.

The reason I like Bladerunner and Do Androids Dream is because they ask if an android (AI) has a soul. Interestingly, I think they come up with opposite answers!

I bring this up because I believe in a soul, a spirit; that humans are more than just personality and memory. If the soul/spirit exists, by default uploading just the personality and memory would result in an artificial representation of the human; an artificial intelligence.


-Pie

[Penforhire]

But... if you can't prove humans have an intangible soul then you're just setting yourself up to be a bigot against uploaded people. Oh, you could easily be right about the soul but I'm thinking of a future when uploaded humans are, or would otherwise be considered, "posthumans."

The plots possibilities are similar to analogous complaints against cloned humans. Would you then call them second-class citizens or no-rights slaves?

Consider, what is the harm if you are right versus if you are wrong? Seems like an asymmetrical risk versus reward ("though shall not kill" and all that).

[EatingPie]

Originally Posted by Penforhire

But... if you can't prove humans have an intangible soul then you're just setting yourself up to be a bigot against uploaded people. Oh, you could easily be right about the soul but I'm thinking of a future when uploaded humans are, or would otherwise be considered, "posthumans."

The plots possibilities are similar to analogous complaints against cloned humans. Would you then call them second-class citizens or no-rights slaves?

Consider, what is the harm if you are right versus if you are wrong? Seems like an asymmetrical risk versus reward ("though shall not kill" and all that).

A cloned human is a human. An uploaded human is not a human... or at best a different human.

There's a whole genre of literature where humans are memory duped. Peter F. Hamilton wrote a good book -- if you know up front you must keep track of like 20,000,000 different characters -- where people wore implants and if they were killed, their bodies were regrown and the most recent dump from the implant was loaded. Thus "immortality" was bestowed on humankind. Pandora's Star was the first part, and I forget the book that made up part 2.

I would still consider this a completely different person, since the breath, the spirit is not bestowed simply on our memories.

I figure if I set the rules now, I can't be called a bigot in the future since -- obviously -- they'll be using my rules!

-Pie