You make a hard point there when you wrote that : The AI is programmed to do something beneficial, but it develops a destructive method for achieving its goal
A robot or more exact AI doens't have the same side as us humans ... let me give you an example:Superb post @stellabelle !
- try to think it like a program ....you press run and input the command resolve the war by using these alfanumeric numbers...01011001010101. The soft or AI will start doing what you said but if we start to think...really : what is the cause of war?
Well the cause is us ! So naturally AI will say that the sollution to the problem is kill all humans.
And you can take more problems like sickness and famine.
For sickness-AI can say destroy all infected subjects
For famine - AI can say destroy all weak and sick subjects and it will be food again. If there is no one who can eat then the food will not get eaten...Logical right if you think it that way!
We are logical beings but olso we like to say that we have fealings to so we canot think like a machine will do and we try to think about the best solution for a problem.
AI doesnt share our point of view and to be onest I dunno if it will ever do.
BUT in my opinion we did not lose the race we can still evolve and improve ...sometimes the turtle can defeat the rabit.
PS :Awesome analogy there between bieber and Vasili Arkhipov that man is trully a hero!
I can't take credit for the analogy. It was part of Max Tegmark's presentation. But his presentation was mind-blowing and it cleared up a lot of my questions. I had no idea that the experts in AI are surprised by the current breakthroughs of the past 5 years. Machine learning is happening now, and most thought that we were like 10 or 15 years away from the AI advancements. That's what I didn't know about....
I am so happy that you too share my entuziasm on this subject, here in my country is 2:35 AM and I canot stop reading your posts . I can't wait to see the next one!
Oh, your enthusiasm is exactly the kind of reason I decided to start writing about such topics! I am very happy that you're liking them. Another one is out today!
You knew about Arkhipov? I had no freaking idea! And my father was even an anti-nuclear weapons person and member of Physicians for Social Responsibility....you do make a good point about how machine logic doesn't take us into account, that's why Hawking and Musk are worried! Also, imagine when the weaponized AI gets so advanced that there is an entire black market of it????????
Also, the other thing few people even consider is that when superintelligence gets to be recursive, meaning that it will be able to improve it's own intelligence, and it surpasses us in every cognitive way, then we will no longer be able to relate to it at all! We will have created something which we do not even understand! This is crazy. And most of the AI scientists believe that by 2060, superintelligence will exist............
There are already robots who are babysitters in Japan.......