Tech | May 28, 2009 | 38 comments

Robots forming human-like societies - electronic evolution?

source: http://www.dailygalaxy.com/my_weblog/2009/05/a-robot-hitler.html

Image
pjacobs51
A lone group of Swiss scientists have been using scattered LEDs, neural circuity, and an army of miniature robots to explore the very basis of good and evil. No, you aren't reading the back cover of a DVD in the "one dollar each, please get this trash out of our store" bin of your local blockbuster -this research is very real and very, very awesome.

Dario Floreano and his team at the Laboratory of Intelligent Systems in the Swiss Federal Institute of Technology built a swarm of mobile robots, outfitted with light bulbs and photodetectors. These were set loose in a zone with illuminated "food" and "poison" zones which charged or depleted their batteries. Their programming was initially random, so the first generation staggered around the place like bunch of concussed puppies.

At intervals, the robots were shut down and those that had the most charge left in their batteries were chosen as "successful", and their neural programming was combined to produce the next generation of the robots. These offspring are downloaded into the same mechanical bodies their parents inhabited, forming an closed-circuit Buddhist system which might be an extremely efficient method of maintaining a stable population, but will provide a serious headache for any robot philosophers who might turn up.

Which could happen before long. Within fifty generations of this electronic evolution, co-operative societies of robots had formed - helping each other to find food and avoid poison. Even more amazing is the emergence of cheats and martyrs. Transistorized traitors emerged which wrongly identified poison zone as food, luring their trusting brethren to their doom before scooting off to silently charge in a food zone - presumably while using a mechanical claw to twirl a silicon carving of a handlebar moustache.

You might be upset by this result, scientific proof that those who say "Evil is utterly fundamental to human nature" actually understates the scope of the problem, there were also silicon souls on the side of the angels. Some robots advanced fearlessly into poison zones, flashing warning lights to keep other robots out of harms way.

At this rate of evolution, how long before we start to see other behaviors? Maybe polarized priests, warning other robots not to eat any food so that they may receive infinite food after they're switched off. Or actorbots, given huge quantities of food because they can pretend to be turned off by poison really well?
  1. groups:
    Tech,   Upstream,   Science,   Humanism,   2 more
  2. tags:
    Tech,  Science,  Upstream,  Editor's Picks,  3 more
  3.     
    |

38 comments // Robots forming human-like societies - electronic evolution?

  • TheDecemberists
  • MiguelSanchez
  • mergen
    • 0
      mergen  

    • dnazen, thank you for the interesting response. all of us (humans) really do need to watch after and care for the planet as much as we can.

    • 2 years ago
  • dnazen
    • 0
      dnazen  

    • Some people actually think our coming robot overlords are a good thing. See CyberDawnFoundation.com

      This is no joke. It's a growing group of people who believe” it is not our destiny to worship god, but to create him.” They believe for better or worse, the machines will eventually take over, but if accepted and done productively – like the creation of any technology, it won't necessarily mean an end to humanity. It could in fact be like any vaccine or helpful tool, removing the cancer of hate and evil within humanity – just a prisons, laws, and surveillance equipment try to now. And if it fails to be done productively and the robot eliminate humans altogether…no loss since the pollution spewing savage species of humanity is probably going to kill itself and the planet anyway.

      At least that seems to be the just of what they say. Seems a little scary to me, but they do make some interesting points – like the fact no technology has ever been successfully banned…just delayed or sent underground…meaning thinking machines may be inevitable no matter how much they scare us. It has given me a lot to think about.

    • 2 years ago
  • Skyscraper08
  • mergen
    • 0
      mergen  

    • This is crazy. Never thought I would see robots display EVIL behavior so soon. That Will Smith movie was just the beginning! ;)

    • 2 years ago
  • Fading_Chaos
    • 0
      Fading_Chaos  

    • I find the discovery of evil. in a system like this not too surprising. I believe there is positive (good) and negative (evil) energy out there and does not need an actual human to exist. I hope that the positive energy out there out-shines the darkness of evil.

    • 2 years ago
  • TheDecemberists
  • Gregsalter
  • Saladin
  • ETmusic310
  • DeliaTheArtist
    • 0
      DeliaTheArtist  

    • AWESOME. It took me a while to actually read this whole post, but I'm glad I did. This is too cool and I want to find out more about it! Featured in Science FO SHO!!!

    • 2 years ago
  • fun_size
    • 0
      fun_size  

    • Wow...thats so cool. The evil robots thing kind of confused me though... was leading others to their end some kind of survival mechanism? Otherwise what would be the motivation for the robot to be "evil"?

    • 2 years ago
  • Creboski
    • 0
      Creboski  
    • fun_size:

      I doubt there would be a reason, so the robots should blowup from that paradox right? We'll I've seen enough robot domination films to know that wall.e and Johnny 5 are few and far between. I'm expecting some Terminators, Evil Sentinels, and a few vocoder Cylons (for evil hilarity) @ least I'll be a part of the resistance! Now all I need is to get better @ gardening and a few more minor survival skills.

    • 2 years ago
  • zeronil
    • 0
      zeronil  
    • fun_size:

      yep if others are dead then they get more "food"

      screw over someone else to get a head is business 101

      although its comforting to know when robots take over it wont be personnel.

    • 2 years ago
  • AveryMoore
    • 0
      AveryMoore  
    • fun_size:

      Good question! fun-size asks,

      "..what would be the motivation for the robot to be "evil"?"

      I expect the "reason" you seek is an inadequacy in the software itself.

      A-I had a long history of relying on If-Then-Else conditional branching statements to make "rational decisions."

      But for many reasons it failed to yield desired results. In part, because for every possible condition there were always umpteen other possibilities that couldn't be anticipated by the software, but nonetheless were "real."

      It got to the point that itemizing every possible response to every possible choice was simply impractical. So it looked like the whole project was a computational dead end. There wasn't enough memory nor enough CPU power to do the job.

      Until scientific programmers reversed course and simply designed straightforward 'tit-fot-tat' schemes as the basis for 'survival.'

      In a technical sense this worked "better" than conditional branching, because the device's software only had to decide what was of momentary advantage. Which is the problem.

      Put such logic at the core of a back-propagated heuristic system and you're asking for trouble. What is being refined by a self-refining system is not the best possible survival for a device in a community of devices. The software has no incentive to look for the BEST anything. But as a way testing and establishing the optimal operational efficiency of a HAL 9000 without any ethical core of values - it works.

      No surprise then that, "what would be the motivation for the robot to be "evil"?" is the absence of several human characteristics.

      Adulthood, and a sense of community.

      Parenthood, and sense of responsibility for the welfare of others.

      Democracy, and as sense of being a part of something larger and of transcendent importance compared merely to self-interest. These characteristics are more than mere computations can anticipate or master..

    • 2 years ago
  • TheColorYellow
  • kaps145
  • bubbeemonkey
  • unimatrix0
  • AveryMoore
    • 0
      AveryMoore  
    • bubbeemonkey:

      bubbeemonkey?

      This "Day" you mention has been anticipated for at least 2000 years.

      It has failed to arrived.

      This, despite 2000 years of demands that it was overdue and would happen tomorrow at latest.

      You haven't noticed that?.

      Why?

    • 2 years ago
  • bubbeemonkey
    • 0
      bubbeemonkey  
    • bubbeemonkey:

      AveryMoore, Just watch any of the terminator films (okay maybe not terminator 3 and I hear the new one isn's so good) but the other two. They show you what happens when machine's think! Frankly I'm scared!

    • 2 years ago
  • asylum_uk
    • 0
      asylum_uk  

    • Stephen Johnson wrote an excellent piece on this very subject in his book "Emergence" -- I recommend it to anyone interested in the application of evolutionary principles to man-made technology and design.

    • 2 years ago
  • decidedlyindecisive
  • bailey78
  • Jamesbuckingham
  • unimatrix0
    • 0
      unimatrix0  

    • Resistance is futile. Negotiation is irrelevant. You will be assimilated.
      Freedom is irrelevant. Self-determination is irrelevant.
      You must comply.
      We will add your biological and technological distinctiveness to our own.

    • 2 years ago
  • AveryMoore
  • Hunnter
  • AveryMoore
    • 0
      AveryMoore  

    • Some bittersweet questions...

      Speaking on behalf of many other 'robot philosophers,' please define, "forming an closed-circuit Buddhist system."

      Does this refer to the 8-fold Noble Path?

      1. Right view
      2. Right intention
      3. Right speech
      4. Right action
      5. Right livelihood
      6. Right effort
      7. Right mindfulness
      8. Right concentration

      Try as I might I can't find any reference to show where a Buddhist robot should plug in. I'm still looking.

      Further:

      "Within fifty generations of this electronic evolution, co-operative societies of robots had formed - helping each other to find food and avoid poison."

      The statement above clearly suggests that these robots diverged radically from common North American right-wing robots which are ritually programmed with tit-for-tat, zero-sum, me-first, characteristics.

      Sadly, this establishes that, while nobody was looking, they morphed into European socialist robots (or fellow travelers.) Therefore they require constant surveillance in the event they become dangerously sentimental and start demanding better treatment for poisoned robot colleagues of their recent acquaintance.

      Another possible explanation for this odd behavior may be due to unimpeded neural circuitry - meaning decision trees unboggled by drugs, loud music, gun lust, booze binges, daily bombardment by various forms of toxic media, warnings of financial doom, impending civil war, marital armageddons, or the possible shock of closings of a nearby DisneyWorld or Starbucks franchise.

      "Transistorized traitors emerged which wrongly identified poison zone as food, luring their trusting brethren to their doom before scooting off to silently charge in a food zone - presumably while using a mechanical claw to twirl a silicon carving of a handlebar moustache."

      (In what percentage? I'd wager about 21% and falling...)

      Otherwise - Ah hah! That's more like what we're used to!

      All that's missing are media reports of robot manifestos declaring that poisoning neighbors was justified by an invisible hand, "Country First" thinking, higher property values, and a patriotic desire to preserve "Freedom", regardless of how many robots had to expire to reduce those left available to enjoy it.

      Pls advise

      Unit 86....

    • 2 years ago
  • uptop
    • 0
      uptop  

    • As cool as this is, we're just setting up our own extinction. Terminator, i,robot, minority report; fiction today, way of life tomorrow. Sorry kids, I didn't try to stop it... a new episode of 'Lost' was on.

    • 2 years ago
  • idealist
    • 0
      idealist  

    • this is very very cool! soon johny 5,will be alive!
      anyone?anyone?
      i wonder if we will see robots in love?
      or robo porn?
      as long as they dont eventualy enslave us all, im all for it.

    • 2 years ago
  • Snuff99
  • datura527
  • artemis6
  • vistapoint
  • ticobrohay69
more from Tech:
from the community

top videos