I would guess closer to 50 years from now myself, maybe 2070. Mostly because science has no idea where consciousness comes from. A programmer can create a neural-net that acts like a brain, but, how do you give it a consciousness and free will?
Your consciousness is not the brain itself, but rather "the watcher" or "the decider", aka "the man behind the curtain"... how do you give such an aspect to a computer?
Some people think that once a neural network becomes large enough it could become self-aware spontaneously. If this happened it could be a lot sooner than 50 years. Though, I don't see how you could even tell it was self-aware unless it had the ability to reprogram itself.
A self-aware AI with the ability to reprogram itself would evolve faster than people could keep up with it. Soon we would not understand it's code, even if it allowed us to view it. This sounds like the "singularity" you refer to. You can't put that genie back in the bottle once it is out...
Your consciousness is not the brain itself, but rather "the watcher" or "the decider", aka "the man behind the curtain"... how do you give such an aspect to a computer?
Some people think that once a neural network becomes large enough it could become self-aware spontaneously. If this happened it could be a lot sooner than 50 years. Though, I don't see how you could even tell it was self-aware unless it had the ability to reprogram itself.
A self-aware AI with the ability to reprogram itself would evolve faster than people could keep up with it. Soon we would not understand it's code, even if it allowed us to view it. This sounds like the "singularity" you refer to. You can't put that genie back in the bottle once it is out...
You are touching directly the important point here, the theme of consciousness and how far we are from creating not artificial intelligence (because it has been already created) but artificial CONSCIOUSNESS, which basically means: A mechanism that is able to reflect on it's own condition and capable of learning and re-organizing itself to overcome its obstacles.
That's the real phylosophycal deal in this field, which has been explored by Phillip K. Dick in his novel called "Do androids dream of electrical sheeps?" and in a Ridley Scott movie called "Blade Runner" which is an adaptation of the mentioned novel.
I like the position that Ben Goertzel (Probably the greatest IA developer in history) takes: He sees consciousness not as a consequence of a complex interconnected web of neurons but as the most fundamental 'substance' by which all of reality is made.
For he, it doesn't depends on if a robot has consciousness or not, because he believes that everything we see, touch and experience is just an expression of (the fundamental) consciousness, just in different degrees and shapes. So everything we can perceive is made of the same consciousness and because of that, EVERYTHING has consciousness. (If you want to listen him talking about this specific topic watch this video at minute 9:12 https://www.youtube.com/watch?v=owppju3jwPE&t=515s
It may sound crazy from a materialistic/traditional science point of view, but even the most advanced scientific understanding that we have nowadays (which is Quantum Mechanics) may prove that there are evidence of it being totally possible.
