Team Aims To Create Pure Evil AI 527
puroresu writes "Scientific American reports on the efforts of Selmer Bringsjord and his team at the Rensselaer Polytechnic Institute, who have been attempting to develop an AI possessed of an interesting character trait: pure evil. From the article, 'He and his research team began developing their computer representation of evil by posing a series of questions beginning with the basics: name, age, sex, etc., and progressing to inquiries about this fictional person's beliefs and motivations. This exercise resulted in "E," a computer character first created in 2005 to meet the criteria of Bringsjord's working definition of evil. Whereas the original E was simply a program designed to respond to questions in a manner consistent with Bringsjord's definition, the researchers have since given E a physical identity: It's a relatively young, white man with short black hair and dark stubble on his face.'"
I have no mouth and I must scream! (Score:3, Interesting)
http://web.archive.org/web/20070227202043/http://www.scifi.com/scifiction/classics/classics_archive/ellison/ellison1.html [archive.org]
Re:How do you define evil? (Score:5, Interesting)
Not necessarily though, I mean don't get me wrong, I think Child molestation is a bad thing, but sometimes the people are so messed up inside that they KNOW its wrong and they can't stop doing it, like how some people can't stop smoking cigarettes.
Having known someone who was into that kind of thing, he told me that he really hated who he was and that it felt a little bit like a bipolar thing that he couldn't help. Was what he doing wrong? Absolutely, and he knew it.
Did he feel he was doing more good then causing harm? No. He turned himself in.
I posted because when I read it I thought "How does one create pure evil when evil is a frame of reference?" So I went to RTFA and just thought that Bringsjord's definition of evil was not exactly what -I- would picture pure evil. I imagined pure evil as that maniac who wants to control the world for his own benefit, at the cost of anyone elses lives or pleasures. My closing comment was that Bingsford's definition of pure evil exists QUITE COMMONLY in the world today.
Re:At what point... (Score:5, Interesting)
Re:How do you define evil? (Score:3, Interesting)
Bringsjord's definition has some interesting presuppositions about human nature, apparently.
It kinda sounds like he thinks "evil" can only be born out of "incoherence" (reasons to commit the morally wrong action) or "misunderstanding" (regarded the harm as a good thing).
It also is interesting that he doesn't define what a morally wrong action is, or what is morally wrong. It seems that is more to the point in defining "evil." If I define "morally wrong" as that which only applies to interactions with others as opposed to being "morally wrong" with interactions that have only to do with myself (say... pride, selfishness, etc., which do not require "others" to exist), my definition of "evil" will be significantly different.
Re:At what point... (Score:5, Interesting)
"I think evil would look fairer and feel fouler."
True evil would try to look as trustworthy and pleasant as possible; or, to also paraphrase Baudelaire,
"The greatest trick the Devil could ever pull would be convincing the world he didn't exist."
Re:How do you define evil? (Score:4, Interesting)
I imagined pure evil as that maniac who wants to control the world for his own benefit, at the cost of anyone elses lives or pleasures.
I thought this too. Then I wondered how you could analyze such an AI. A big part of being pure evil includes deception with lies and half-truthes. One would almost need two ways to interact with the AI: one as a random person and one as the "always gets the truth" person.
Re:The Scary Door from "The Spanish Fry" (Score:3, Interesting)
Former Student (Score:3, Interesting)
Re:Pure Evil? Check out latest contract killing. (Score:3, Interesting)
But, a cold-blooded murderer is not an excellent personification of evil. They have their reasons for doing something, usually something twisted or self motivated.
True personification of evil would not be a murderer, because he or she would evaluate their actions and consequences. If I kill this person, I will end up in prison, and eventually executed. It's hard to continue your evil streak if you're locked away or dead.
It could be argued that authors are closer to evil, as they can write novels based on countless heartless killings, yet they've never been caught nor convicted for. In that, they've planned and reviewed the entire action of murder and the consequences.
Most cited murderers had a purpose of some sort, either in reality or their own twisted fantasy world. It may have miscalculated results. There was some gain to it. It could be for themselves or someone close to them. That in itself indicates some level of good. Consider the cold war strategy of "If we're nuked, we nuke them", held by both sides. The initial aggressor would be attempting a conquest for the good of their country. The retaliatory strike would then defeat the aggressor. The outcome? All of humanity is destroyed. (ya, ya, I know, not enough nukes, but it paints the picture).
I'd love to play in a virtual environment with E. I would likely show their developers that their "Evil" isn't as bad as it could get. Annoying? Sadistic? Maybe, but the human mind can always create deeper darker levels of evil than anything that can be programmed.
Re:How do you define evil? (Score:4, Interesting)
Disintegrating a populated orphanage with high explosives for fun is not evil, its psychotic.
Disintegrating a populated orphanage with high explosives because you truly and firmly believe that the world is better off without those orphans, and then convincing the world to see it from your point of view, and getting away with it.
*THAT* is pure evil.
Selmer is great (Score:1, Interesting)
I went to RPI in the late 80s when Selmer Bringsjord had just joined as a philosophy professor and taught classes in logic and philosophy and coumpter AI. I wrote a computer simulation of the board game Risk that ran on a symbolics lisp machine for one of my projects. ah good times...
anyway, Selmer has a knack for bringing attention to his projects -- in addition to this evil ai system I remember he was very interested in a system which could generate television scripts, story plots, etc. based on interactions between story elements and simple intermingling of them. The idea got wings and humanities people were upset because they saw it as a way to replace writers and artists with computers. http://kryten.mm.rpi.edu/brutus.preface.pdf
Re:Pure Evil? Check out latest contract killing. (Score:4, Interesting)
There is a bit of finesse here. Lets presume for a second that Obama's health care plan would actually benefit some segment of the population (real people, not corporations, government or HMOs), opposing it on purely inhuman ideological grounds could be seen as evil, if we accept human well being as the ends for all good actions. He's forcing something on us, and purely ideological opponents are also forcing something on us. Its the act of imposing your will on others which is the common point.
We run into a problem with calling Obama's health care plan evil though... A majority of people voted for him, knowing EXACTLY that he would do this, and a majority of state voters voted for the majority party in congress, and should have known that this would happen. I take this as acceptance, or at best complacence. We (as in Americans, not us individually) wanted this, and thus it isn't being forced on us.
Yes, there is problems with this, but these problems are rife in any democracy (a republic being a form of democracy). The tyranny of the masses is built into the system, a little "evil" will always leak in, but this is arguably better than the alternatives (one person or group inflicting their will upon you).
The other problem with ideological opposition (notice the word "ideological") is that it ignore the real world, and human consequences. When you oppose something that effects humans for non-human reasons, then you are doing nothing but trying to inflict your ideology on others as well. Lets say, for instance, that you are a strong libertarian, this is fine, as long as you don't think that this is the only way of seeing things, or the only valid way. Your ideology must be balanced by the ideologies of other interested parties, only through that compromise do we minimize the evil of imposing our will upon others.
I, for example, have some heavy socialist leanings (not in the common misuse of the term as ad hominem and partisan smear), but I would NEVER want to live in a purely socialist country. I like the idea, but realize that it fails at several levels, several of which are going to be completely opaque to me. I need the moderation of opposing views to correct the flaws in my own mental schemas. Ideologies, in other words, exist in a void, detached from the human consequences of their imposition into reality.
If you oppose Obama's health care for real, human related, reasons then there is no evil there, as long as you can acknowledge that there are people with equally valid views on the other end. I as a socialist REALLY dislike his plan as well, to remove any partisan element from this. If you oppose it just because you have an ideology, then you are just as bad as other people inflicting ideologies on you. Just because you agree with it, doesn't make it good or just.