Attack of the Killer Robots

I wrote a post a month or so ago, which I called Humanity 2.0, which was quite widely circulated and worth a read if you missed it, even though I say so myself. My basic premise was that as Artificial Intelligence take us nearer to The Singularity (where machines take over the world), we should think and debate about what that new world might be like. Because there are some very scary scenarios that might emerge.

This week, a bunch of eminent scientists agreed with me – though I’m sure they didn’t read my post – as they expressed concern over the increasing use of robots in combat roles.

The problem….was that robots could not fulfil two of the basic tenets of warfare: discriminating friend from foe, and “proportionality”, determining a reasonable amount of force to gain a given military advantage.

While the headline to this post sounds like the title of a bad B movie from the 1950s, it is an area that we need to start deeply thinking about. Rogue robots are one of the more obvious manifestations of machines taking over. I’d argue that we need to also be concerned about the more subtle and innocuous use of machines in decision-making capacities and the implications of things going wrong, whether that be due to malicious intent, viruses, hackers or even the potentially sinister “machines know what’s best for man”.

Do mankind a favour and share the Singularity vision with a friend. Help them understand the consequences. If we’re going to make this happen, let us at least understand the implications for our future.

—–>Follow us on Twitter too: @russellbuckley and @caaarlo

Share this post:
  • Facebook
  • Twitter
  • del.icio.us
  • StumbleUpon
  • Digg
  • LinkedIn
  • Google Bookmarks
  • Suggest to Techmeme via Twitter
  • Reddit
  • Fascinating... scary and got me thinking...

    I see UAV/drones as the interim technology; and they have already been used to kill people - http://www.airforce-technology...

    'In November 2002 in Yemen, a Predator UAV was used to drop a Hellfire missile which destroyed a civilian vehicle carrying suspected terrorists.'

    Somebody, thousands of miles away, "suspected" the civilian vehicle contained terrorists... I guess we'll never know if he/she was correct!

    Then I wondered how much info the pilot of these drones have - from http://science.howstuffworks.c...

    'Predator aviators have described piloting the aircraft as flying an airplane while looking through a straw. This is quite a change from driving a conventional aircraft from the cockpit.'

    I think most of us agree that if we were driving a car trying to get our family away from a battle zone we'd like the pilot to have more than a 'view through a straw' before he/she vaporised our loved ones.

    Can the desk-bound pilot thousands of miles away be charged with war crimes?

    It seems to me that we have interim questions to answer before we get to AI and autonomous drones.
blog comments powered by Disqus