• Question: will AI become complex enough to become self aware?

    Asked by hello1237 to Vince on 17 Jun 2013.
    • Photo: Vince Hall

      Vince Hall answered on 17 Jun 2013:

      That is a very difficult question to answer, and I don’t think anyone can know that unless they are the AI itself.
      Nevertheless, if you follow the trend in the growth of power of computer hardware and software you should see that the power increases exponentially. This means it doubles every time. What time? Well for some things it’s a year, so the number of processes a computer processor can do in one second seems to double every year for the same price. So £1000 a year from now should buy you about twice the processing power of £1000 today. Laptop computers today can do billions of processes in one second, my phone does 1.5 billion a second. Billion = 1000 million = 1000 1000 1000 = 10^9.

      The complexity and sophistication of programs is increasing exponentially as well. It is said by a researcher and entrepreneur called Ray Kurzweil that computers will reach human levels of intelligence around about the middle of the 2020s. He says, however that by the mid 2020s, you should be able to buy a computer with human-level intelligence for $1000!
      This remains to be seen. Don’t believe everything you read, many others say it’s really much slower, but he has studied this a lot, so he might be right.

      Ray goes on to say that non-biological (computer) intelligence will be billions of times more powerful than all humans by about the 2040s. Cool!

      I would assume that at some point they would reach self-aware status if their intelligence increases exponentially. That is very different from just pure computing power.
      This should be very good for us, as they would be able to do tones of complicated stuff we don’t, and can’t understand.
      Some hope to make AI part of their personal intelligence; then they could control it properly.