Here is an interesting article about Elon Musk’s opinion on AI (Artificial Intelligence).
Being a Sci-Fi Writer and a Aeronautical Engineer with a MS in Computer Technology Management, this article really hits home. I totally agree and I’ve had tons of experience with computers doing exactly what he talked about. Computers do what they are programmed to do, but a complex program always takes on a character that may be different from what the programmer intended. In many instances we had programs that we thought were ready for release and invariably they would do unpredictable things when faced with real world input.
Carrying this problem into a future world where there are computers with AI, unpredictable results are guaranteed. Coupling those results with a computer program that can learn and adjust itself will be like letting the ‘horse out of the barn.’ The key element to be sensitive to, is giving the computers the connections and ability to ‘do things.’ If a program gives the answer to ‘eliminate humanity,’ the programmer should be able to intercede.
Applying a little Sci-Fi thought to this, I can see a world where a computer with AI is built and over time it does a good job and finally someone thinks about how much money they can save by trusting the computer and giving it control. After all the typical management thought is, ‘if it hasn’t broken, then convince me it will break,’ which in this instance is saying ‘if I can save money, then we’ll turn on the computer and if you can’t prove that it will fail, then I’ll assume it won’t,’ and of course the implied ‘yes, but, if it does fail then it’s your fault for not programming it better.’
In my series SIMPOC, my computers follow this kind of logic we end up with a good computer and a bad computer. The bad computer does what it is programmed to do. And “Yes,” there are aliens and a apocalyptic virus thrown in for fun.