RE: (SteemSTEM) Machine Learning and Learning Machines: Any Sufficiently Advanced Technology is Indistinguishable from a Brain
This is quite interesting, but also confir.ation of my greatest fears.
A.I will surpass humanity. And if humanity were to be destroyed by AIs, it wouldn't be because of any superiority complex, but rather a logical conclusion that it would be better off without us.
Your "getting warm" argument holds true.
AI has no moral obligation or mental blockage that prevents it from thinking all the possible solutions to a delema .
The extra thought humans put to processing whether a thing "should" be done is unknown to an AI.
The ending of humanity comes when AIs can effectively repair, fix, create and destroy other AIs.
Once the loop is self sufficient without human input, the next logical solution would be to get rid of any threat to their programming .
Heck, a directive such as "protect humans" can and will be solved the most efficient way, which may not be in favor of humans but goes perfectly in line with said directive.
We are doomed if we so slightly mismanage this ai revolution.
I'm not so much worried about intelligences smarter than us. That might even end up being our salvation. I am worried about algorithms vastly dumber than us but more numerous, that interact with our tools and homes and vehicles, reacting to each other in sneaky and/or unpredictable ways. Driverless cars getting hijacked by advertisment software, excessive wifi signals causing drones to crash, facial recognition going horribly wrong, advertisements with eye-tracking capability that won't play unless you are looking at them ... stuff like that, only stranger and worse.
@edumurphy, great article as always bro, love the arguments brought up here
@edumurphy, great article as always bro, love the arguments brought up here
@destinysaid You bring up a very interesting point that technologists and futurists are thinking about today about whether future AI will have goals that align with our agenda as humanity.
But before you can ask that question, you have to ask the preceding question of just how superior AI can become? Is it possible to reach an AGI or ASI? Are there any upper limits on the development of these neural networks?
And can these AI Neural Networks become not just "doing machines", "learning machines" & "analyzing machines" but can they become "conscious machines"? Currently we are able to program AI within certain confines of our own goals. Is it possible to create truly autonomous AI that can create its own goals? We know that AI is able to achieve amazing feats and goals but the question is can they set their own goals and think on their own or have a mind of their own? And if so, would they feel the need for setting their own goals, would they feel rewarded for achieving their own goals? A fascinating discussion that AI brings, it is just mind-bending and mind-boggling to see how fast technology keeps escalating today! We are literally thinking about the psychology of neural networks (large repositories of code) and how software can become conscious! Just think about how crazy this discussion that we are having is lol!
Very crazy indeed sir.
For me, the questions pertaining to AIs and their capabilities isnt one that has a simple answer.
Conscious machines ? Who knows, at what leavel can we say a machine has achieved consciousness?
What is consciousness ? Can it be measured ? Etc
The old saying holds true, the more answers you get, the more questions you generate .
We solve one problem just to find 4 more behind it.
Frankly , a conscious AI would be a very good thing, it can be reasoned with.
Ill check your blog out for more of these sort of info, im not too well informed about AIs, but its a subject ive watched and grown fearful of over the years .
That's true. Thank you, I'll check you out as well.
I love reading, learning and engaging with the community on AI! I haven't written too much about the topic yet, as my primary focus has been on crypto reviews, but you can check out this post I made a while back which discusses AI in the context of what the world will look like 20 years from now.