AI and their motives

Seems like it would have been easier, not to say cheaper, just to bring along some more humans. Even fully sapient (do you mean sentient?) AI's would likely not be really human.

I mentioned that humanity wanted at least another form of Sapience to be there.

Humanity is weird, considering simple warp drive and jump gates are in setting AI didn't seem too far off for me.

So there's no "Frankenstein Complex" in your universe? Here in the real world, if you so much as mention AI everyone is all "Oh no Skynet!"

There was a time that was a worry. It passed. Time and experience has changed the general perception, it helps that few AI in recorded history have gone crapshoot.

Some even see the tales of A.I.s going mad as Values Dissonance.

edited 22nd Feb '15 5:14:09 AM by EchoingSilence

Cool.

This train of logic leads to a world made of paperclips.

Heh. But that was my point, actually.

You need to define harmless very carefully. An AI not designed to place humans and their well being above all other objectives will probably eventually be lead to do things that are harmful to people.

To quote:

My point is that AI's dont "decide" to become harmful. Terminator style aggression isnt very realistic. Either the dangerous potential is present in the original programming (presumably by accident) or it isnt going to happen.

Possibly. If it's a tightly designed AI, with no emergent possibilities, then yes - it will have a set of parameters it won't deviate from, simulating a certain deisgn of thought process.

But if we create an intelligence that isn't bound by human norms and is emergent, it COULD, if it can learn, develop hostility, just as it COULD develop empathy and understanding. I agree, Skynet is unlikely in terms of an AI deciding... but look at the quality of software being pushed out the door now. A hostile AI could very well happen due to shitty bug testing.

The irony that QA would be responsible for the end of the world.

Hence why in every universe I have ever made AI research takes great lengths to prevent psychopaths that appear because they are A.I.s.

If some go mad it is more human style madness and wanting to harm someone and you can never tell who it is until it happens.

Divine Automata in my world are divine robots granted with sentience and sapience by their owners. They rely on both massive amounts of divine circuitry to run their processes, and an "Artificial Soul" which gives them their human-like qualities.

DA's never look human: trying to emulate humanity completely, all the way to having skin and the like, is impossible and makes the process fail. Instead, DA's are humanoid, they just have "skin" made from some non-organic material. Think Alchemical Exalted. Depending on who made them, they're wrought from divinely-strengthened metal, carved from gemstone, made from starmatter, or are hewn from putrescent flesh.

They are made by a variety of people to be bodyguards and assistants. Despite being AI at their core, they are nearly indistinguishable from humans emotionally, possessing emotions, likes, dislikes, dreams, hopes, fears, and the whole package. They also have sexual desire, empathy, and sympathy. Truthfully, sometimes their artificial exteriors are the only indicator of their non-human nature. They harbor no ill will towards humans, glad to fight alongside, work with, or have sex with their masters. They are not bound by their masters, however, and pissing off your DA is a good way to get a chainsaw up the butt and a free DA on the loose. Most societies in this world don't mind DA's, but they are viewed as servants. A DA who strikes out and lives on their own would be viewed as a quirky weirdo at best or an insane rouge robot at worst.

DA's tend to have complexes over their dual nature: they have human-like souls and sapience, but it comes from masses of divine circuitry and a soul crafted by human hands. They are robotic, but they don't have the blinding soulless efficiency of computers. How they choose to deal with this is up to them.

A decent amount of DA's practice sorcery, being called "Divine Processors" and their spells called "Heaven-Cracking Programs". Being robots, they weave these spells by hijacking the area they want to affect through sorcerous means, using their robotic minds to crack into the "code" that underpins reality, and then altering that code to warp reality.

edited 3rd Mar '15 1:54:20 PM by Frostav

The "main" AI in my setting is a hyper-intelligent computer who views the entire world as a broken system in need of repair. To this end, he managed to find a few human sycophants, hacked a few important systems, and ultimately created a nation (a constitutional republic of sorts, where his "coding serves as the constitution". A supreme court trial amounts to asking him and waiting around for him to punch out an answer). Civilization is a machine, that humans live inside and are a part of. The good news is that the AI falls into The Extremist Was Right territory very often. In fact, he's more or less the Big Good of the (rather cynical) setting, and is very instrumental to humanity's continued survival. Because he's a machine not programmed for selfish behavior or emotion, he's more or less incapable of abusing his power (or even truly understanding why he'd want to). He is capable of being wrong, though, if he's given incomplete or incorrect information, or becoming too much of a Knight Templar.

I've got a new post apocalyptic setting and in it, AI aren't crapshoot because they are AI but because someone put in some malicious coding because they didn't like humanity. This setting is a bit bleaker and more cynical compared to my other settings which always have a shred of hope and the possibility of improvement.

Hell going Crapshoot is akin to getting a Virus. It could happen to any digital entity and considering some of them hold super powerful bodies, like the protagonist who is a android himself.

That being said for the most parts A.I.s are again seen as not too different from humans mentally.

Protagonist: who built them and why?

The Protagonist, which I call Spade, was built as a Core Guard for the Center of the empire, well what was the empire for humanity in setting. His prime directive was to protect and serve but since interaction and questioning to find out what the hell needs to be protected and who shall be served was a big thing (as the empire had an entire council going) Spade was made to be so human that very few people could tell otherwise.

A few decades after the Empire Crashed, humanity divided it into small empires in what they controlled of the Galaxy have kind of eroded that prime directive. Spade's on a long trip back to Earth, and that prime directive has sort of manifested as a Chronic Hero Syndrome.

Hell a lot of A.I.s in setting changed a lot over time, some becoming more robotic others becoming more human. Some with influence some with none.

God (or something similar) built him, in an act of divine delegation. The planet of the setting is basically a giant climate-controlled biosphere. A few A.I.s were put in charge of monitoring the planet (IE, creating new forms of wildlife, controlling the climate, shaping the inhabitant's technological advancement, ect). The other A.I.s broke down or went insane by this point (he has an evil counterpart who's basically an AM-expy).

In my own settings, A.I.s aren't coded or written - it proved i possible to create something sentient with code. Instead, volunteers have their neural patterning (re: their concious mind) transferred to clones... or, often, a quantum matrix.

This essentially produces the quick-thinking, highly-capable aspect of an AI, but with a personality and without the whole "Re-written to kill everybody" thing. The few who have gone rogue have done so because existing mental disorders (in one case, psychopathy) carried over; an AI of any kind has no motive to go on a murder spree.

For the purposes of the setting, yes, it is the actual person who's been through a Brain Uploading; not a copy. Whether that works in real life is up for grabs.

Wait, God designed your AI's? Calling them "artificial" seems odd somehow.

Yeah divine intelligence seems better.

I do think I only have one AI that is Crapshoot and not insane or malicious by human standards and that's only because it's alien in origin and runs off of Blue-and-Orange Morality so what it sees what is doing (IE assimilation into it's own core directive everything around it) as good. This is a relatively new idea actually.

edited 15th Mar '15 4:29:11 PM by EchoingSilence

Technically the term would still apply to any intelligence designed by another intelligence.

That would include us.

That is a good point. He is an AI in the non-literal sense of being a sentient computer.

Yes DeMarquis, We are The AI.

Dun Dun DUUUUN!

edited 15th Mar '15 9:45:56 PM by Ekuran

"You will have no choice but to enjoy phase 2 of our program..."