Skip to main content

Home/ READ 0880 SP12/ Group items tagged Intelligence

Rss Feed Group items tagged

ryanhersha

Control AI before it controls us, expert says - Technology & science - Innovation - msn... - 2 views

  • doomsday scenarios could be prevented if humans can create a virtual prison to contain artificial intelligence before it grows dangerously self-aware.
    • ryanhersha
       
      The dangers of self-aware AI might be contained if artificial intelligence were contained in a kind of "prison."
  • One starting solution might trap the artificial intelligence, or AI, inside a "virtual machine" running inside a computer's typical operating system — an existing process that adds security by limiting the AI's access to its host computer's software and hardware.
    • ryanhersha
       
      The AI "prison" would be located within a "virtual machine" separated somehow from the host computer's other workings, the internet, and people.
  • "If such software manages to self-improve to levels significantly beyond human-level intelligence, the type of damage it can do is truly beyond our ability to predict or fully comprehend," Yampolskiy said.
  • ...11 more annotations...
  • "It can discover new attack pathways, launch sophisticated social-engineering attacks and re-use existing hardware components in unforeseen ways," Yampolskiy said. "Such software is not limited to infecting computers and networks — it can also attack human psyches, bribe, blackmail and brainwash those who come in contact with it."
    • ryanhersha
       
      Super-intelligent computers could develop powers we don't presently associate with computers.
  • proverbial
  • cryptography
    • ryanhersha
       
      def.: "the practice and study of techniques for secure communication in the presence of third parties"
  • oracle
    • ryanhersha
       
      def.: "A priest or priestess acting as a medium through whom advice or prophecy was sought from the gods in classical antiquity."
  • but without the sure knowledge that any of the steps would really work.
    • ryanhersha
       
      This radical uncertainty seems to be a defining characteristic of discussions around the consequences of AI developments. Given this, might experimentation with super-intelligence might be inherently irresponsible?
  • That would harness the power of AI as a super-intelligent
    • ryanhersha
       
      Oracles hold power.
  • slow down the AI's "thinking"
    • ryanhersha
       
      & if we're too slow one time...?
  • mere human-level intelligence could escape from an "AI Box" scenario
    • ryanhersha
       
      even humans have broken out of most prisons. the difference with AI might be that the escapee could be super-human
  • "The Catch-22 is that until we have fully developed superintelligent AI we can't fully test our ideas,
    • ryanhersha
       
      Do the ideas' untestability render them less visible, less serious in the academy? We have no problem developing new technologies, but controlling them or predicting their impacts is another matter.
    • ryanhersha
       
      I have highlighted and sticky-noted this article as an example of how to complete the "Sticky Stories" assignment. You don't have to highlight using special colors, but so you can see an example of each, the blue highlights are for vocabulary, the yellow for main ideas, and the green for my "connections."
  • Never send a human to guard a machine
    • ryanhersha
       
      Humans guard machines? My iPhone already tracks my every move.
1 - 1 of 1
Showing 20 items per page