In theory, given enough computational power (more then we currently have) an algorithmic program (in an environment we don't currently have) able to self replicate and mutate could give rise to true AI, but it would not be english speaking humanistic intelligence as portrayed in movies/books. I see something like a quantum computer virus being the path to AI.
It is by caffeine alone I set my mind in motion. It is by the beans of Java that thoughts acquire speed, the hands acquire shakes, the shakes become a warning.
-Kujako-
.
"This will be a fight against overwhelming odds from which survival cannot be expected. We will do what damage we can."
-- Capt. Copeland
Well if you consider yourself in 20 years, AI Researcher would likely end up closer to a study of AI behaviour than research into improving AI. . . Human beings are rapidly going to need insight into WHY AI think the way they do once theyre smart enough to go beyond the algorithms that spawned them. We already have more than enough 'DESTROY ALL AI!' on the internet that some mutual understanding would probably be pretty important when the time comes....
I would advise that person to pursue whatever job they wanted. The first introduction of A.I. is not going to drastically transform our economy or what jobs are performed by human beings. That shift will not happen until the deployment of A.I. is streamlined and available to a more generalized population. When it is first developed A.I. will be the plaything of research scientists, governments and the military.
As A.I. becomes more closely integrated with other markets and avenues of employment it will be commiserate with arguments protecting "Human" jobs and likely mandate having human beings who oversee the work of whatever the A.I. is tasked with. It will be a long time even after the development of A.I. before your job is in jeopardy.
Anything that require dexterity or talents in some way as artistic will be safe. Even something as simple as a chef will be hard to replace by a robot, his tools will get better, be the chef will still be there.
Another thing overlooked is there's a tipping point where AI start taking on more and more of the research into AI. I'm not sure how far that is off, but even a fairly simple AI can be a great help in developing the next generation of AI, once that trickle of Ai development turns into a torrent humans will be quietly shoved out of the way.
.
"This will be a fight against overwhelming odds from which survival cannot be expected. We will do what damage we can."
-- Capt. Copeland
Being self-aware and intelligent doesn't meant it's fast at information processing. For all we know, current computers are pathetically slow at certain tasks such as learning, pattern recognition, deduction, induction so on and so forth even if you use a neural system. I assume the AI you are talking about is a neural network of some sort. Human visual system, for example, is known for processing massive amount of data in a very short time. As far as I know (don't quote me on that), human brain can perform many petaFLOPS, that's rival to supercomputers. However, the AI you are talking about needs to be implementable to a regular PC. We will not have super laptops 20 in years, that's pretty much granted. So even if we solve the puzzles of conscious and self-awareness or intelligence in general, they may not be as smart as us due to processing speed limitations. Many people do not consider these issues. An example is Neural Networks are old stuff, most of the theoretical framework was well-established around late 70s and early 80s, however, they were not utilized for long time because of hardware limitations.
Another problem is training problem. Neural systems require training and the more neuron you have the more training you need. Humans require decades of training (living, we call it) to reach maximal mental capability.
Last edited by Kuntantee; 2015-09-02 at 06:21 PM.