What is feasible with AI and what is not?

Hello

I have a assignment that I really need help with. I have used google and reddit but cannot seem to answer the above question in lay man terms.

I would be very grateful if someone could please give me an explanation of a generic understanding of what is feasible with AI and what is not.

many thanks

2 Answers

Relevance
  • 1 month ago
    Favourite answer

    Well, it's not an easy question to answer. We're talking about artificially creating intelligence... the ability to learn from experience, and think independently.

    Science has so far enabled this in a limited way, in very specific situations.

    There are, for example, chess programs that can get better and better at the game by referencing previous games and referencing the statistical likelihood of a sequence of moves. The best chess computers these days can beat any human, they have an unfair advantage in terms of how far ahead they can calculate moves, and knowledge from thousands of top-level chess games.

    https://www.extremetech.com/extreme/196554-a-new-c...

    But are they intelligent? Not in the sense that a human is.

    Can we make a computer brain that does all of the same things that a human brain does? No. We're not even close to that. The human brain is incredibly complex, and many things about it are still beyond our understanding. Perhaps one day, we will understand the brain well enough to build one. But that's a long road, with many complications attached.

    What if we get it wrong? What if we create an artificial intelligence that is insane or psychotic? How do you create compassion, and fear, and hope, and love and hate, how do you create emotion? If we ever do that, would that intelligence be "alive"? These are questions that many science fiction movies have posed to us (like the synthetics in "Alien", or the replicants in "Blade Runner", and there are countless other examples... Westworld, I Robot, various Star Trek episodes...).

    We have artificial intelligences that can pretend to be a person online, and do it fairly convincingly for several minutes (Google "AI conversation online" and you will find something like that. Or look up "Eviebot"). We have Phone-assistants like Siri and Cortana that are quite clever at predicting our wishes. But these are algorithms that make statistical predictions, we are a long long way away from recreating truly independent thought.

    And yet, we are close enough to realise that we may get there some day, create true artificial intelligence. Close enough that top scientists have expressed concern about our research into it, such as Stephen Hawking, before he recently died.

    Like I said, it's a complex question, you could write a book on this, and people have. Hopefully some of this is useful to you.

  • Bryce
    Lv 7
    1 month ago

    Computers, like IBM's Watson and Wikipedia, are able to store vast amounts of information, but they are unable to synthesize it and form brilliant ideas to solve complex problems and create great new products.

Still have questions? Get answers by asking now.