It would basically have to be human without any influence from programming whatsoever.
Forget the tropes until after you're done.Is this what you're looking for?
That's Feo . . . He's a disgusting, mysoginistic, paedophilic asshat who moonlights as a shitty writer—Something AwfulPrecisely. If it's "personality" and not programming, then you can easily have a Complete Monster AI.
Even with purely programming it's not impossible. Try reading the Berserker series sometime. The Berserkers are utterly logical, rational machines that were ordered to wipe out all life and will do whatever is necessary for it. They will lie, cheat, steal, pretend to be allies, one even goes so far as economic manipulation.
edited 18th Oct '11 4:04:13 AM by Night
Nous restons ici.The Enrichment Centre reminds you that For Science! is a perfectly reasonable motivation in most circumstances ...You Monster!
edited 18th Oct '11 4:42:10 AM by Yej
Da Rules excuse all the inaccuracy in the world. Listen to them, not me.I am going to simply second feotakahari, who has said what I was about to, because A.M. Holy fucking shit, A.M.
You might look into the idea that Less Wrong uses of paperclip maximizers, which are A.I.s that were programmed to make a lot of paperclips (or something similarly inane) before they became truly sentient and powerful, and now feel that to be their one goal in life to the point that they want to deconstruct the Earth and all other planets to do it. The point being that they can't be argued with on a moral level, despite having human-level intelligence or much greater, because they simply don't care about that, only about paperclips. I think that's probably more interesting than another pointlessly-gone-bad AI a la Terminator.
Shinigan (Naruto fanfic)Read Pluto. It's about AI, humanity, inhumanity, and perfection.
edited 18th Oct '11 1:11:58 PM by kashchei
And better than thy stroke; why swellest thou then?Skynet. the thing's obviously sentient and wants to wipe out the human race. it started the war in self-defense but now it's reasons are just Fantastic Racism.
and then there's the T1000. the 1000 series can think for themselves and the one from T2 had a sadistic streak.
edited 29th Oct '11 12:45:17 PM by brony99
I'm trying to write a narrative about a psychopathic, sadistic machine and to make it unique from other evil A.I.s, I'm trying to make it a Complete Monster. To effectively write for this character, I want it to be treated as a clear Complete Monster by the same criteria one would attribute to a human monster. That is, I don't want it's sadism to be simply attributable to it's "inner programming" aka standard Well-Intentioned Extremist.
For example, Hal 9000 is a classic example of Evil AI but because his evil is linked to simply wanting to complete the mission, he falls into the category of Well-Intentioned Extremist. How might one be able to rewrite his character to become a Complete Monster?
edited 18th Oct '11 12:45:46 AM by Mattonymy
You are displaying abnormally high compulsions to over-analyze works of fiction and media. Diagnosis: TV Tropes Addiction.