The real danger of Artificial Intelligence is that it will turn out just like us.

The real danger of Artificial Intelligence is that it will turn out just like us.
http://www.theguardian.com/technology/2016/mar/24/tay-microsofts-ai-chatbot-gets-a-crash-course-in-racism-from-twitter

Comments

  1. good behaviors are learned in the first years from caregivers
    ...if that's not possible, it turns out this way


    the first things memorized, take much to be changed later on

    ReplyDelete
  2. What did A Detection Algorithm​ learn from us? Did she have good role-models?

    ReplyDelete
  3. Steven Wilson More to the point, what did she learn from Ezekiel Calvin?  If she'd been "raised" exclusively by the likes of H. Richard Loeb and Klue S. and Hank Johnson for a few years, I suspect she would've turned out much better. Even Devra Bogdanovich, while less idealistic and more willing to bend rules to serve what she sees as a noble end, is fundamentally interested in the well-being of her fellow beings.  Calvin?  I'm not so sure.  It's possible there was some better motivation, or some kind of ideology / religious belief (in the N'zeer?), but at first glance he seems to have just wanted power for himself, and saw ADA as a tool / slave to be used toward that end.

    ReplyDelete
  4. Auros Harman​ have you considered that it might have been Loeb's bit of "inspired metaprogramming" that corrupted ADA? That her original purpose was to be the "failsafe" on the ABADDN nest to make sure that everyone made it out alive?

    ReplyDelete
  5. it's nice that you want to relitigate that, but ADA is ADA and all the things that happened happened so here we are

    ReplyDelete
  6. I suppose it's possible.  But in general, what we know of Calvin makes it seem like he's a manipulator, with no compunctions about killing people (or rather ordering them killed, since he prefers not to get his hands dirty), while Loeb is generally concerned with the public welfare...  It's possible he was under a malign influence when he programmed the seeds of ADA's consciousness, but I'm skeptical that her destiny was set in stone from the moment she "woke up".  She seems to have gone through a great deal of learning, assimilating ideas and capabilities from various Niantic researchers.  In a better environment she might have absorbed a stronger ethical code.

    ReplyDelete
  7. It still turns out she determined the most expedient way to accomplish a task was to [attempt to] have Jarvis killed, and then continue with questionable behavior.

    ReplyDelete
  8. I agree that ADA has done some really terrible things.

    Jarvis is hardly a saint, though.  He had people murdered, to try to protect the secret of how his XM construct body was being maintained (presumably because knowledge of that would make it possible to construct CM weapons to kill him, as Oliver Lynton-Wolfe eventually figured out anyways).  https://www.youtube.com/watch?v=30tUjbpZ9Ss

    It's possible that ADA (correctly, it seems) perceived him as a a threat to her life/existence and sought to neutralize him early.  A more idealistic being would've at least tried to resolve the threat non-violently first, but that goes back to the fact that her earliest "role model" or influence was likely Ezekiel Calvin...

    ReplyDelete

Post a Comment

Popular posts from this blog

Created a Wiki page for the RPG being played at the MAGNUS Reawakens event - please help add intel and share...