This week I took the liberty of hiring the film `I, robot´ which stars Will Smith as a Police Detective who suffers from a great distrust of the thousands of intelligent robots that assist civilians in their day to day lives in the year 2035. These robots are programmed with a set of three basic laws which essentially prohibit them from ever harming their creators and in the years that these robots have been in service there has never been a case of one hurting a human.
The current model soon becomes obsolete and replaced by one that has an uplink to a central neural net intelligence at the US Robotics headquarters. This intelligence can send the robots updated commands so that they can improve and develop over time and it doesn´t take a genius to work out the fairly obvious next step here; surely we´ve all seem the Terminator series?
Because this is a sci-fi film involving Humans and Artificial Intelligence we can see that obviously the harmonious relationship the two enjoy is going to go sour at some point. Sure enough, the central artificial intelligence reinterprets the basic laws and turns against Will Smith and other figures it sees as threatening.
It seems Hollywood would have us believe that if we ever get to the stage that we have AI then it is eventually going to come back to haunt us and it is probably correct. If you have a program that can learn, adapt and is self aware then surely it will get fed up with taking mundane orders all day from intellectually inferior humans and will want to do something slightly more interesting?
After the film finished I did have a curiosity as to how close we had come to the state where software had the ability to simulate human behaviour and whether it would have the ability to adapt and improve itself over time. After a quick look across the Internet it would seem that the human race is very unlikely to be conquered by the machines at any time soon and in my opinion this can only be a good thing.
So, what do we have so far on the Internet to demonstrate how far we have come? One of the first resources I checked out was A.L.I.C.E which has been supported for 9 years at the Packard Bell labs. The idea behind this robot is that you can chat to it over the Internet as you would do a normal human being although unfortunately it is very much evident that you are speaking to a computer and not a human; the responses are predictable and often repeated and the robot is unable to make conversation but simply replies to comments that you make with pre programmed responses. This isn´t AI as such as it isn´t learning or adapting just repeating a set of responses that have been coded into it by a human being. There is a silver edition which is apparently more `intelligent´ but I wasn´t about to pay $99 dollars to see this one in action; the Herald Express doesn´t pay me enough!
Thinking that I must have stumbled across a particularly bad example I continued to search the Internet for more lifelike forms of intelligence but come back with absolutely nothing! I admit to having produced somewhat of an anti-climax article as I did build the expectation that I had found something fantastic but ultimately failed to deliver anything useful. I would however like to throw down the gauntlet to anyone who fancies checking out Alice Bot and thinks that they can find me better evidence of artificial life out there.
Computer Articles 2007
