Will A.I. doom the human race?

Here is another great article on artificial intelligence.

Stephen Hawking: Artificial Intelligence Could End Human Race

http://www.livescience.com/48972-stephen-hawking-artificial-intelligence-threat.html?cmpid=558754

I agree with the concerns this article presents although it is a matter of interpretation and I suspect that the BBC and the person who wrote the article enlarged the concern rather that expressing it clearly. After all you get more readers if you interpret the comments with GIANT RAMIFICATIONS.

I haven’t been able to read the actual comments Mr. Hawking said so I can’t comment on those.

I worked a great deal with computers during my career and I was responsible for huge systems. I agree there is some risk with AI but I think responsible controls are all that’s necessary. I think the risk is less than other risks posed before us.

AI could run amuck if the computer was self-learning, which we have a little of now. It would also be a concern if it could modify its own programming; which is a higher form of self-learning. But for the computer to be a true risk it must control its power supply and have a mechanism to manipulate its environment. Both of these are easily controlled by a human.

Acting like a human is much different than thinking like a human. Computers can be made to search giant data bases but they make decisions based on objective data instead of subjective data. If the result of their search is 12, they will make a decision based on that result. Building in the mechanism to do more than that is the challenge. One of our great strengths is the ability to use our past experiences to modify our decisions. The computer power to accomplish that is huge and will take computers a long time to replicate.

True if AI had the power to modify its programming and manipulate its environment we might be in trouble. After all it could go through 10,000 generations of evolution in the blink of an eye.

In my series SIMPOC I’ve tried to replicate that point in time where computers can think, manipulate their environment and act to defend themselves. I picked that point in time based on current research predicting that point should occur in 40-50 years.

It will be interesting to see how computer technologies change and given Moore’s Law who knows it might happen faster rather that sooner.

Even in SIMPOC I have the ALL POWERFUL ‘OFF’ SWITCH.

 

Ray Jay Perreault

http://rayjayperreault.wordpress.com

SIMPOC Series link:  http://amzn.com/B00KCAB6L0