So I was reading through Scientific American when I came across an interesting artcile. In Troy, NY Selmer Bringsjord is trying to discover the root of all evil (which I always thought was the love of money). He is a philosopher, logician and the chairman of Rensselaer Polytechnic Institute‘s Department of Cognitive Science. He has decided to create a team of graduate students to simulate an evil person on the web. Now as Sci-Fi has warned us, this often results in the rise of machines but this man does not fear the results.

One of the hypothesis of what makes someone evil is that they seek to do something morally objectionable to harm others without any provocation. If said person is asked to give a logical reason as to why they wanted to cause harm, their reason would either be “nonsense” and also reveal that the person knew what they did was wrong yet still had the desire to do it.

So one must wonder, how do you make an evil program? Well it becomes an even greater task of how do you make a program think for itself? Let’s take Bowser from Super Mario fame. He was evil, but only because he was programmed to be evil. He was following a script and didn’t have the capacity to do any form or reasoning to lead him to do anything evil. All he could do was steal a princess and spew fireballs at an Italian plumber. He couldn’t send you emails  as a prince from Nigeria who needs help in order to access his fortune and will gladly pay you for your assistance resulting in your bank account being emptied.

We have not seen true AI yet. We have a few small versions of AI but we haven’t seen a computer program with that ability to think and reason and emote like a human being. So how do they do it? First they have to start by coding into it an identity. The basics that go into the identity are things like gender, age, name, etc. and from there progress into giving them a “background” complete with motivations, history, beliefs, etc.

So now they have a man named “E” (for evil, hah!) who was first built in 2005. At first he just responded to questions that fitted the original hypothesis of “what is evil”. Now he has evolved into a digital person with facial features which I think makes him look suspiciously like  Sylar, the primary antagonist from NBCs Heroes.

“E” currently dwells in a digital world and they have even had him interviewed by a digital Barbara Walters (not really but wouldn’t that be fun?). During this interview, they used a scenario mapped in psychiatrist and author M. Scott Peck‘s 1983 book, People of the Lie, The Hope for Healing Human Evil. In this scenario a parent gives a boy a gun which is later used by his older brother to commit suicide. They programmed “E” to believe that he gave the boy the gun. They then asked a series of questions to figure out why he gave him the gun. “E” told him that he had a gun, the boy wanted a gun, so he gave the gun to the boy.

So what is the end all goal of this? To better understand human behavior. We can currently study humans and how thy think and react but we have yet to actually jump into the brain and see all of it functions. Well now they are essentially creating a mind and hopefully will be able to figure out what causes a person to be evil. Now obviously this raises some interesting ethical questions. What if “E” was dropped into Second Life and started killing off random avatars? It may be interesting if he were dropped in World of Warcraft, given a warhammer, and just see what happens. Obviously, there is no interest in dropping “E” into any kind of virtual environment without safeguards. They will follow Isaac Asimov’s “Three Laws of Robotics” which essentially dictates that a robot is a subordinate to humans and must obey and protect them.

I personally would not be shocked if they created a virtual environment full of “E’s called SkyNet and then it takes over the global network and Judgement Day takes place. Of course, we would then discover time travel which would be cool. I honestly think “E” should be dropped in Final Fantasy 14 as the final boss. That would make for an awesome video game. Imagine having a bad guy who just does what he wants, doesn’t follow any script. Imagine if Bowser just started throwing Goombas at Mario on Level 1 for no reason or just destroyed level 2 so you couldn’t progress? That would definitely increase the challenge.

Anyway, it’ll be interesting to track the development of “E” and other forms of AI. It will also be interesting to see what would happen if pure evil could be calculated and digitized. Who knows what we will unlock in the technological future.

Categories: Torchlight News ~ Trackback