AI requires electricity. We, not AI, control electricity. There are all sorts of fail safe protocols that could be legislated to be required of all AI manufacturers, and made part of international trade agreements. I don't see it ever becoming an unmanageable problem.
"Sorry Dave, we can't allow you to do that today". Fun stuff for Hollywood, but just isn't reality. I was amused by 2001, but I saw it to be a hole in the script. If we were smart enough to make Hal, we would have been smart enough to make fail safe non Hal bypassable kill switch/safety protocols.
If AI can self-learn, then if it goes rogue and its main power source is compromised, don't you think it would look for alternate forms/sources of power? Something else to consider...
We only control electricity because only we can build generators. For now. As automation increases (which AI is directly related to), we control the production and distribution of electricity less and less. Also the advancement of nenewables decreases our control over electricty (which is one of its big selling points). AI wouldn't have to be 'free' for long to, for example, steal a credit card number, buy a server, satelite internet subscription and a buttload of solar panels and 'go off the grid.' We're not 'there' yet, of course, but we may not be very far from it either. Don't make the same mistake of every govt that has fallen in history by thinking that because we control things now, we will thus always control them. Part of the allure of AI is the opportunities for automation that it represents in the utilities sector. If we sufficiently automate power (or water, or food distribution), AI could end up controlling pretty much everything, then its just a matter of how careful we are about not allowing backdoors to exist in the 'off switch' we'll surely build into it. And apparently we humans love to create backdoors that we arent sposed to. due to our non-commital nature and desire to leave our options open for future opportunities. No doubt, we could build a careful system to coral AI and prevent it from going out of control. But it will require a level of vigilance and perpetual dedication that we have not demonstrated to be able to keep.
Evolution isn't a 'designer', it's a process. We, who design AI, are not a process, we are designers In short, it's an apples/oranges comparison.
I see no evidence of it. Can you elaborate with real life and well known examples? I see no evidence of it. Care to elaborate? any electronic device can be disabled with high powered electronic beams of some kind. when we get that far, we will have a disabling weapon. A Ray Gun, so to speak. Well, what you are imagining is science fiction. Yes, there is the old saying, because there are many examples proving that yesterday's science fiction is today's science fact, therefore, today's science fiction is tomorrow's science fact. However, since not all yesterday's science fiction is today's science fact, therefore, not all today's science fiction will be tomorrow's science fact. What you are contending is possible, but I'm not seeing it, so far. The big reason I don't believe it because if it were possible, these machines would advance to conquer other planets, and given 14 billion years of our universe's history, it should have happened by now (alien machine conquerors) and since they haven't conquered this planet, I am doubtful that your science fiction will ever become science fact.
How do you know it hasn't? Since what we're discussing is very much science fiction (so far as we know), an interesting movie that depicts precisely what you're talking about is 'Moonfall.' In case you like science fiction movies, I won't spoil it for you (any more than I already have, anyway). Its reasonably entertaining, in a very 'that's not how that really works' sort of way. But you seem to be assuming that AI would have been invented somewhere near the beginning of the universe. I don't see why... First something has to evolve to build AI. Then AI has to evolve. Then AI has to spread. Then AI has to travel between the stars. Not necessarily difficult task for machinery, but still potentially an incredibly time consuming task. Perhaps we're building the first one. Perhaps it already wiped out all life except for a few 'cockroaches' that fell in some bubbling ooze here a few billion years ago, and we havn't popped up on its radar yet. Or anything in between. Hell, maybe we have it all wrong and biological life is actually an evolutionary result of sufficiently advanced mechanical life. I mean, ya, its all pretty far fetched, but so was space travel until we did it. 50 years before that, most people had never even seen a car. I don't think you or I are under threat of being squished by giant sentient robots. But perhaps in a few generations. I am all for leaving them a nice stockpile your 'EMP beams.' Get on that.
I didn't say that....(more on this, below) That only is logical from an anthropomorphic viewpoint. The problem is there is no 'alien' in 'anthropomorphic', so it's not something you can actually and thus factually claim. whose to say what aliens and/or alien robots more advanced than us can achieve? Maybe they found a way to traverse intergalactically in a short period of time. I don't think so. Anyway.... Assuming infinity, all that is possible is inevitable. I believe that in all sincerity, that given infinity (assuming it) anything that is possible is inevitable. It's just a matter of time. As far as we know, the universe is what, 14 or so billion years old? That's not infinity, but close enough, in my opinion, that the above statement still holds true. Yes, it's possible, that something could be possible which didn't occur within 14 billion years, but that is long enough time that if anything is possible, it will most probably and plausibly occur within that time span, and account for evolutionary stages. Also, let's factor in quantum computing. Toss that in there, and the more likely the equation is true. "In my opinion" (which is all anyone has, including that of Scientists). I didn't say that [it happened somewhere near the beginning of the universe], but I would suspect that if something could occur, factoring in evolutionary time, noting that when quantum computing level is achieved, development accelerates exponentially, there is no telling what will result, over time, but if something could occur, it would have happened by now. So, if, say, it was developed by some race 6 billion years ago, on another planet, and, similar to earth, it took the planet 4.5 billion years to make the first quantum computer, and 40 years later, AI took over, and quantum computing advanced several fold, that leaves 1.5 billion years after taking over, to develop. What could an AI dominant world achieve in 1.5 billion years? Galactic propagation? It's entirely possible, and probable, in my view. Seems to me my idea that 'assuming infinity, all that is possible is inevitable' such that, if it hasn't happened by now, it's not likely to happen/ Either that, or they just haven't discovered us yet, and that I do not believe, because, if you read the many books on the subject, by Prof. John Mack, Budd Hopkins, and Dr. David Jacobs, humans are being abducted by aliens. They have found us. But are they sentient beings or are they AI? Maybe they are AI. No way to know. A very compelling book, entitled "Walking Among Us' by Dr. David Jacobs. You might want to check it out.
Whatever this technology gets used for, those in "high places" will do it before we know it and despite anything we might think.
And it's a process that AI developers are using a lot. You are obviously not included in that, "we." People have even designed computer viruses to evolve, let alone AIs. Only if you don't know how modern AI and AI development actually work.
The same could be said of biological intelligence, but where are they? Moreover, unlike biological intelligence, superhuman AI won't necessarily have a motive to "conquer other planets." Until we have evidence of life and intelligent life on other planets, we have to face the possibilities that either life itself is a wildly unlikely fluke that has never emerged anywhere else in the universe, or that technological intelligence is a wildly unlikely fluke that has never emerged anywhere else.
That is false. That's not how it works. 14G is not only not infinity, it is not even what any mathematician would call a large number. False. You have opinions. Science has facts. That is pure supposition with no basis in fact. Wrong. "If." Puh-leeeze...
I'm not an expert, true. But I do know that AI uses electricity, which is something mankind can control, if he is smart. However, if AI can access the theoretical 'zero point energy' without the aid of man, then you might have a point. But when I say 'you would have a point', one must factor in that I'm not an expert, and it's a loosely conceived opinion.
A very compelling book, entitled "Walking Among Us' by Dr. David Jacobs. You might want to check it out.
All numbers are small compared to infinity, of course. It's all about the IF. No one really knows. Read the book anyway. I suspect you are supposing it's all 'woo', but it's really not at all. Dr. Jacobs is not a dummy.
Until we give it control of the electricity. The management of an electrical grid requires constant attention and adjustment. A job AI is tailor made for. Allowing AI to control the grid would potentially save energy companies millions of dollars. It is eventually going to happen.
WEll, I, as CEO of a company, could give you, manager of a department, authority to act autonomously, but I would reserve the power to fire you if you became a liability to the company. Metaphorically speaking, why would it be any different with AI?
