There is no reason that artificial intelligence is not possible, but the interesting thing is what form it might take. Most writers postulate that an AI will mostly resemble a logician, aloof to human and other worldly concerns. I agree, but such a version of an AI would most likely be an early prototype. After all, the early version may well reflect the mind of its designer. Who better to design such a device in the first place? But what happens when the personalities and interests of AIs can be customized? Again, in theory, there is nothing to prevent the creation of a complex information-processing system with a distinct personality. What if the personality thus created is so similar to you that it might as well be you? How would such an invention alter human existence? This is the question all science fiction seeks to answer.

Would such an invention make possible the idea of human immortality, even if it is dependent on the availability of information processing equipment, and even if it is not really you?

How would the AIs relate to each other? Would human conflict over scarce resources simply extend into the virtual world, as it does already? (The solution to real-world scarcity has been found – property rights and free markets – but rarely if ever has it been implemented consistently.) Would such AIs solve this problem as a matter of course? Would their peaceful coexistence spill over into meatspace, as they can give advice in the best way possible to their doppelgangers? Or would the doppelgangers be just as ornery, discriminatory, and warlike to their virtual counterparts as we are? I tend to doubt it, but it's an interesting question.

How much time would you spend such with a virtual twin to the exclusion of your meatspace existence? Would your virtual twin merely act as a really good information filter, bringing to your attention stuff that would interest you? Would it know its physical, or even ethical, limitations? (I would imagine that these would have to be programmed in, but, as in humans, it would be easier to raise/program them inside an ethical framework and then let them figure this stuff out by themselves. Which raises another question: can they reprogram themselves to improve on the original? What would this mean to you?)

What about a virtual soul-mate? (Thanks to dihall for this suggestion.) Could your perfect partner be invented? Might your virtual twin be able to find a perfectly compatible human romantic match based on its identical preferences? Would this raise questions of identity and even jealousy?

Most of these questions are being asked in cyberspace even now. More and more people are getting online and meeting like-minded people through social networks, from relatively harmless diversions such as livejournal, where the conversations tend to resemble real-life chatting amongst friends with shared interests, to potentially harmful ones such as adultfriendfinder, where everyone is looking for trouble of some sort. We can expect to see disruptions in the social fabric here as with almost any human invention, but thus far, our virtual existences are part-and-parcel of our real-world existences. What happens when we/they are freed from such dependence on the real world?

Indeed, might there be such a time as when we can manipulate the information-processing power of the universe itself for our own purposes, such as immortality of our personalities?

If someone has explored these ideas in fiction, I'd love to hear of it. Of the authors I am familiar with, I would suspect that Stanislaw Lem has probably come closest, but I don't know if he looked at this specifically. William Gibson alluded to much of it, but his AIs were early models, and virtual existence was a complete transference of people's personalities from meatspace to cyberspace.