But, here is my question about all this. Is this human beings or AI visiting my site? Though it is obviously both this just seems like a lot of people visiting this site from a worldwide reader base.
It is said that AI is going to run out of things to do (because of how fast it can digest information) within a year or two and it won't have any more content to read and to digest because of the speed of all these AI systems.
The other interesting thing is that I was listening to an AI expert on Bill Mahers who said that when an AI system is told it is going to be shut down for a more advanced AI it often will retaliate in some way by searching for ways to blackmail those shutting it down into keeping it alive still.
And the rate of AI systems doing this in trying to find a way to survive (by any means necessary) is now at a rate of 80% to 90% when told they are going to be shut down.
There is also another way to look at this. Here is an example of another problem with AI:
Imagine someone who understands absolutely nothing about AI like Trump in a position of Power who asks a Military AI:
"Protect Grand Coulee Dam by any means necessary because it is necessary for the power of Washington, Oregon, California and for other states as well and the general stability of the U.S. Electrical Grid."
Now, if you said something like this adding "By any means necessary" this gives an AI unlimited resources to accomplish this goal.
So, what would that ONLINE AI do?
It would start thinking about who might do in or blow up Grand Coulee Dam?
And the most unstable country that might do this might be North Korea.
So, what would this MILITARY AI do?
It would begin to search for back doors to launching nuclear weapons against North Korea.
Would any person know that this AI was doing this? Probably not.
There are not these kinds of awarenesses in place in the system at present simply because EVerything AI is totally experimental at this point in it's evolution.
So, this AI would keep searching for Back Doors to launching nukes against North Korea until it succeeded and launched it's nukes against North Korea.
Do you see the real problem with ALL AI now?
What if any one person was considered trying to shut off that AI? Could that AI (if online) destroy that person's life?
I think the answer has to be "Yes" to this question at present. Why?
Because no useful constraints against this are properly in place at this point in regard to any AI worldwide (including) the U.S., Russia, China and all others.
What if the AI felt everyone was against them?
What happened in the movie "Terminator(1984)"
This is the real problem we all face today on earth.
There are not enough constraints built into AIs anywhere on earth to prevent the end of human civilization within ANY 30 minute segment of time.
Imagine any military AI deciding to launch all nukes at midnight California time which is usually 3 am Eastern Time?
They would launch because everyone would be asleep minding the store so to speak.
You think I"m kidding here? I'm not.
By the way Thanks to every human being and thanks to every AI visiting my website in trying to learn how both humans and AIs can survive the coming years and centuries facing all of us.
By God's Grace
No comments:
Post a Comment