5

Philosophers have discussed the question of personal identity of humans, including humans uploaded to hypothetical computers; but are there any notable discussions of the problems of personal identity of a future strong artificial intelligence (outside the context of uploaded minds)?

  • Would not the same problems exists if the ai was truly ai? Is artificial intelligence not the mere replication of the human experience through the digital medium as opposed to a biological one? – user2683 Jul 16 '15 at 17:16
1

Philosophers have discussed the question of personal identity of humans, including humans uploaded to hypothetical computers; but are there any notable discussions of the problems of personal identity of a future strong artificial intelligence (outside the context of uploaded minds)?

I'm not a philosopher; I'm doing research on general cognitive theory which includes both human and machine cognition.

  • Computers are based on Church-Turing theory. This theory does include AI but does not include things like understanding and reasoning which humans do everyday.
  • Human-level comprehension and reasoning requires a different theory from Church-Turing and hardware that is different from a computer.
  • Even with such hardware, it would be impossible to use a machine like this to transfer or even duplicate your mind.
  • With a general theory of cognition it should be possible to build a machine that is conscious with similar comprehension and reasoning to a human.
  • Since such a machine would be self-aware, you are correct that this could lead to psychological problems for the machine agent.
  • 1
    Please elaborate. Why do you claim that "Human-level comprehension and reasoning requires a different theory from Church-Turing"? What kind of general theory of cognition are you talking of? would it be possible to model such a future machine mathematically? suppose we build it. would it not follow the rules of quantum mechanics? if it does, then what should prevent (in principle, not in practice) a Turing machine from simulating it to arbitrary accuracy? Can you provide some details on your research and institute? – nir Apr 8 '18 at 20:19
  • I have evidence for it. This is my field of research. A theory that would explain how human reasoning evolved from unintelligent round worms to humans, that would explain how human reasoning works and how it could be replicated non-biologically. If the theory is completed then a human-level machine could be designed. I have no idea why you are talking about quantum mechanics. It hasn't had any effect on the research I've done. There is nothing in Church-Turing that allows for defining or replicating human understanding. So, how would you add that? Magic? – scientious Apr 9 '18 at 9:22
  • My research. I created knowledge theory which provides a foundation for cognitive theory. I can explain awareness, learning, understanding, and problem solving. The theory covers things like worry and attention blindness. It isn't finished though. If I do finish it by 2019 then I'll probably publish it along with knowledge theory in 2021. – scientious Apr 9 '18 at 9:22
  • Turing does not deal with thinking nor understanding simply because there was no and there still is no accepted definition for them. That is the point of his paper "Computing Machinery and Intelligence". I don't understand, are you proposing a way to build machines that can not be simulated even in principle by Turing machines? i.e. machines that are not Turing equivalent? – nir Apr 9 '18 at 13:11
  • @scientious, are you an academic working in a field with peer-review practices, in which you plan to publish the theory? – elliot svensson Apr 9 '18 at 16:06
-3

A human is a whimsical being: sometimes it can act against even it's own mental programming. From what we can gather, it is impossible to set a straight rule to what a human does or feels, and for a computer to try to simulate that would be like making it choose a random number from 1~10, except that each number has equal probability: but the human choice does not. The human also acts based on it's feelings: a computer could be programmed to respond to certain feelings that can be triggered by some sort of observation/emotional simulation program that is fit to the target human's mental/physical state, but those feelings won't be real. It may seem like an oddball feeling, but the reality of what someone feels can make a world of a difference. For example, imagine your mother telling you that she loves you. You know she means it, you know she loves you with all her heart. But take away the realness, the sincerity of those words. What are they now? Even if the words come to your ears, they wouldn't touch your heart. Humans are emotional creatures. As much as we are lying, thieving, cold, heartless, ruthless, selfish, conceited, self-serving, spoiled, cowardly creatures, we also have an unbelievable love for the truth, at least from the emotions of another significant being. Computers can't simulate truth: only physical demonstration. This will block the computer from being aware of the identity that they were uploaded to simulate: they may physically act like the identity, but they won't emotionally be able to contain the complexity of the human brain. Of course, this is talking about the first two or three decades of the 21st century. Maybe computers will develop new hardware to make itself notice it's uploaded identity and perhaps even contain the human soul, but for now, I don't think it's possible.

But hey, that's just my theory.

  • 1
    philosophy of personal identity is about something completely different - take a look here: plato.stanford.edu/entries/identity-personal – nir Jul 16 '15 at 4:52
  • Welcome to Philosophy.SE! A few bits of feedback: (1) the community style is generally not about presenting things as "my theory is X", (2), the long block of text is challenging to read, can you try breaking up into paragraphs? – James Kingsbery Jul 16 '15 at 17:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.