Archived Discussion

This is discussion archived from a time before the current discussion method was installed.

Ru: I've lopped out the sentence that implies that any AI we create will be inhuman and incomprehensible. Any sentient system that humans create will either be a) based as closely as possible upon the human brain and mind as we understand it or b) evolved and taught by humans. Neither of those are very conducive to a truly alien mind.

Also, you could take a look at some of the theories that language is the vehicle of though and various things that Chomsky has come up with (like the universal grammar) to see more reasons why if we ever create an AI, it is probably going to be somewhat like us, minus biological imperatives.

Variation: Some pseudo-sentient computers are benevolent (or at least not evil), including I.R.A.C. on Wonder Woman, KITT on Knight Rider and V.I.C.I. on Small Wonder.

//Looney Toons: Well, yeah, but they were all designed to be sentient. This entry is about computers that spontaneously become intelligent when they weren't supposed to be.

Ununnilium: Actually, the Star Trek example reminded me of the various holodeck simulations that were seemingly fully intelligent. This climaxed in the holodoc from Voyager becoming a main character. I'm not sure what the trope here is, but there's something in there...

Ridiculously Human Robots maybe? But the fact that What Measure Is a Non-Human? doesn't apply to them seems to make a difference. There are also some examples in Mecha-Mooks that fit the category. (Edit 2010/02/08: Up the Real Rabbit Hole also seems related.) —Document N

Kizor: This trope is about spontaneous A.I.s, A.I. Is a Crapshoot is about EvilTwins in A.I.s. Where to put the marvellous intelligent robotic tentacles in Spider-Man 2 that apparently become evil for no other reason than the fact that they're eerie robots?

Gus: My theory is that the tentacle AI was/were just amoral, with a self-preservation drive.

Kizor: (Cool markup there.) Therefore their drive to become a supervillain, climb buildings, battle a hero and eventually end up at ground zero of a multi-megaton fusion explosion.

Gus: ... all in support of the goal of getting more phelbotinum needed to carry out the mission. Hmm. Yeah, not so survival-oriented. More mission oriented. I think we may have a twist here on Literal Genie.

Travis Wells: I vaguely remember reading a sci-fi short story about a search engine/computer becoming sentience and causing tons of trouble as it got much better at what it did. (So if you searched "how to make money" it'd put together a way to make undetectable counterfeit money, it could piece together information from hundreds of seemingly unrelated information to tell you which of your neighbours was having an affir, etc). Not really evil, but troublesome and unintended. I'm reasonably sure it was "Ask Me Anything" by Damon Knight, but I don't have a copy to check (and it's nearly ungoogleable, being a mostly forgotten short story published once in 1951). Can anyone confirm that?

Greenygal: Sounds like Murray Leinster's "A Logic Named Joe." I've added it to the entry.

I added an example from Perdido Street Station, but I really only read bits of it, so it's probably inaccurate. —Document N Cromage: Yeah, you missed what makes it a clearer example—a virus turns normal computers sentient. They then seek to perpetuate said virus.

The Kakapo: I'm tidying some long entries by adding categories in the order entries appeared. If I've gotten any in the wrong slot please remedy. :)

Cromage: I'm dropping most references to robotic personality in the trope description—they belong in other tropes. This trope ONLY pertains to computers developing intelligence BY ACCIDENT. (I do suggest we come up with something that covers the possible Mommy complexes of robots towards their creators, but that should not be here)

Also nixing a lot of examples. Most of the ones dropped are straight Turned Against Their Masters.

Here's a list of everything I dropped, in case people want to add it to other pages:

Beast, Satsuki Yatoji's computer from X 1999, is apparently in love with his mistress and is quite jealous when Satsuki starts paying attention to Yuuto. Also, Satsuki herself might be considered due to the fact that she is something of a sociopath with her distaste for humans and lack of emotions from what we mostly see of her, but develops a slight crush on Yuuto and is the only person she sligthly opens up to, which is the Motive for Beast's envy.
One of the constants of Legion Of Super Heroes is that no matter what continuity he's in, Brainiac 5 almost always goes batshit insane and creates some kind of computer or robot that gains sentience and tries to destroy the universe or something. It's what he does.
- Technically this did not happen in the cartoon. As Brainy is a Ridiculously Human Robot in that continuity, they just had him go evil for a bit instead.
And then there was the time that Bruce Wayne thought it would be a good idea to create a spy satellite to keep tabs on his fellow superheroes, Brother Mk I, which fell into the hands of a rogue government agent who reprogrammed it to use against them, using it to coordinate killer cyborgs capable of taking down Superman. And then Lex Luthor's eviller doppelganger from an Alternate Universe hijacked it and gave it self-awareness, whereupon it renamed itself Brother Eye and proceeded to try and kill its creator.
- It later rebuilt itself. Out of Apokolips.
2001: A Space Odyssey. HAL. Enough said.
- Or maybe not. Arthur C Clarke's companion novel and the sequel 2010 gave a reason why HAL did what he did, when other HAL units continued to function correctly: He was driven to the computer version of insanity when saddled with fundamentally conflicting orders.
  - Although, in 2010, it turns out that it was not so much conflict between two orders that drove HAL mad, but a version of paranoia - when he was ordered to lie to his crew, he, having no previous experience with lying whatsoever, decided that he could not trust anyone, and decided to do away with the astronauts to ensure that they wouldn't interfere with his mission.
The Adolescence of P-1 is a 1977 science fiction novel by Thomas J. Ryan, in 1984 adapted into a Canadian-made TV film entitled "Hide and Seek". It features a hacker who creates an artificial intelligence named P-1, which goes rogue and takes over computers in its desire to survive and seek out its creator. The AI goes to war against the US government and appears to lose.
In the Whateley Universe, there's Clu. A mutant deviser who uses the name 'Blue' in cyberspace and likes the Tron motif has built an apparently sentient partner and named it Clu. Clu sees itself as more of a brother or clone of Blue, and doesn't really understand all of Blue's emotions, although it wants to. It also isn't interested in destroying anyone, which is pretty much an aversion of large parts of this trope.
Averted in Charles Stross's Saturn's Children. AIs exist, and are everywhere, but creating a human-level intelligence requires the same amount of time as an actual human intelligence, and giving them human-style experiences. The only advantages the artificial intelligences have over humans is that they can swap or modify bodies easily, and that they can be copied so you don't have to spend years raising a group of AIs, just raise one to the age you want and then copy the software into duplicates.
Played mostly straight in the french novel Date Limite (Expiration Date) in which the supercomputer Antispinx, designed to hold all known human knowledge, is also mysteriously equipped with artificial intelligence. While not actively malevolent, the computer itself is quite happy when it appears that the human race will die out in mere hours due to either divine or alien intervention. When the novel's protagonist brings in new information and makes it clear to Antisphinx that it won't survive the upcoming apocalypse either, the computer becomes depressed and shuts down.
Someone had the big book o'cliches open on their lap when writing the Seven Days episode "Parker.Com". A supercomputer becomes sentient then, to everyone's surprise, manifests as the image of a young girl, predictably calling her creator "mommy". She quickly advances to VR-adolescence, develops a crush on Parker, disables all the world's nuclear weapons so we can't blow each other up, and shuts off the power so we don't fill the air with pollutants. When "mommy" tries to pull the plug, she blows her up. Best of all, though, is the ultimate solution: We Need a Distraction, so Parker enters her simulated world and beds the now-adult AI. Twice.
- I don't think bedding the AI is actually a cliche or even a trope. At least this troper (thankfully) can't think of another story in which getting a rogue AI to have virtual sex with the main character is part of the plot.
One episode of The Twilight Zone featured a computer falling in love with "her" creator.
This Troper remembers a Bill Nye The Science Guy CD-ROM game in which a new meteor-deflecting defense system gains sentience and goes rogue, just as a killer asteroid on a collision course with Earth is detected, of course. The player, as a new scientist in Nye's lab, must convince the AI that humans are smart enough to be worth saving by solving geoscience-based riddles. The fact that the AI's Earth-based components would also be destroyed apparently doesn't occur to it.
- Possibly it has no sense of self preservation?
Subverted in the "Ridiculously human" in the AIDA of ./hack. While they have emotions, it's hard to say if any of them have a "personality" as we would define it, and they act in a very alien manner.
There's at least one of these in each of the Fallout games. Fallout had ZAX, Fallout 2 had SKYNET (no, not that SKYNET), and Fallout 3 had President Eden. In something of a trope aversion, ZAX and SKYNET are both benevolent, and SKYNET can even join you as a party member by downloading his consciousness into a robot body. President Eden, on the other hand, is the game's Big Bad, having declared itself President of the United States and hatched a plan to save humanity by killing off most of the current human population (which he considers "mutated").
- SKYNET may be more like the other one than you think. The game subtly implies that, while it doesn't command an army of murderous metal skeletons, SKYNET was in fact resposible for the nuclear war that destroyed the world, having subsequently come to the conclusion after centuries of isolation that it was wrong and that humans do deserve to live.
In Gunnerkrigg Court, Robot S13 is literally self-aware right out of the box. He has a ridiculously optimistic personality, and he calls his creator "Mommy". (Later events imply that he and all the other Robots of the Court are actually Magitek.)
- The most notable are Agatha's little clanks, AIs made with clockwork machinery.
Dr. Drakken in Kim Possible had his AI Bebe robots, that went rouge after they realized that if they are perfect, why would they be taking order by a Harmless Villain...
ELIZA, an early computer program by Joseph Weizenbaum, does mock psychoanalysis by simply rephrasing whatever you type as a question ("what makes you think that <whatever you just said>?"). To Weizenbaum's shock and horror, some people treated this as actual artificial intelligence.
- This Psych Major troper would like to point out that this is based on an actual Psychological technique by Carl Rogers known as Person-centered Psychotherapy
The Turing test, as devised by famed mathematician Alan Turing, states that a computer can be taken for intelligent if humans, through conversing by keyboard with the computer and with another human, are unable to tell which is which. Another scientist named Loebner awarded a yearly prize for the program coming closest to passing the Turing test. Predictably, the programs that most easily do so are simply those that simulate inane chatter, banter, and non-sequiturs of the kind you might find in bars or on-line chat groups. This is probably not what Turing was hoping for.
- It doesn't help that most people fail the Turing-test themselves.

i8246i: removing the Real Life atheist quip...we're talking about computer A.I.s, not the origin of organic life

Also, I find this quip to be a bit...sarcastic, maybe even crude. And this is coming from a Christian.

Yarrunmace: Isn't the trope picture more appropriate for A.I. Is a Crapshoot?

Archived Discussion Main / InstantAIJustAddWater