View Single Post
  #10    
Old October 14th, 2013, 04:40 AM
Plumpyfoof's Avatar
Plumpyfoof
 
Join Date: Apr 2011
Location: Melbourne, Australia
Age: 20
Gender: Male
Nature: Sassy
Send a message via Skype™ to Plumpyfoof
Quote:
Originally Posted by BlahISuck View Post
On what basis are you superior, and how is it real? And if AI develop the ability to make complex decisions that aren't as "automated", wouldn't they be developing free will?
I really hate coming off like the super villain in this. I have him in a situation where I can moderate and influence his free will at my whim. I tell him to sit he wants to sit, because I want him to sit. It's similar in that you would consider your president or the king to be superior. They have authority which is respected. See he has placed me on the pedestal of "master" and therefore there is an unspoken yet respected agreement giving me a higher status. I am superior in our relationship.

I want to say yes. But you can't have a forever evolving code run rampant without having basic restrictions and still expect it to be 'moral' (from human standards). So you would place an original code of restrictions that cannot be changed or removed to give it this quality. That piece of code effectively dictates the actions of the AI and so the free-ness of the robot's will is up for debate.


Quote:
Originally Posted by Rezilia View Post
I have a concept, I call the Pandora Code - basically, it's a bit of programming that, once activated, bursts out into (as represented) an ever-evolving universe, ever shifting and changing. This is meant to represent the ability for this "code of thought" to be chaotic - not to where decisions don't have reason but, rather, so that ideas and orders can shift and change due to the robot's interaction with reality. In other words, it gives them free will. Also, due to its ever-changing nature, it is impossible (in theory) to hack or crack. In fact, any time in which someone does manage to hack into it and can be successful, the program will detect it, implode itself, and burst out, resetting the universe to the way it was but having completely different mechanisms which make up the fabric of that universe. To put it differently, code made of always-changeable fractals of code which do not completely destroy the memory or other parts of the program while still annihilating any chance of outside forces getting into it.

In other words, it's the one true 100% foolproof AI program.

Under that program, the merit of an AI robot would be based on themselves, not the "next best upgrade" - so, leadership among them would be determined by the individual - and their ability to reason and exist in society will be equal to our own.
Chaotic is key here. Although it looks good on paper, the chaos will be incomprehendible. If you have any experience coding you will know about bugs and glitches - undesirable side-effects from your code. Because you aren't writing out all the abstract concepts manually you're basically just linking bug after bug together in a semi-structured format.
So yeah, I think when it first starts encountering problems on its own it will attempt to code the most 'logical' response. Logical in inverted commas because the robot has no other experience nor instincts and therefore its logic is highly unlikely to be accurate. I'm willing to bet the program would burn out before it picks up much pace.
Your theory also implies that the code will be able to erase and re-write sections of code to better organise itself and make faster connections and whatever, but have you ever changed a section of code? It influences everything. So it would be forever expanding becoming more complex requiring more and more power to run, even through simple tasks. Unless you have an infinite power source for this thing, assuming it got traction, it's going to run out of juice faster than an iPhone.

I firmly believe true AI is out of the question with current technology. Once quantum computing becomes a commercially available thing we'll be in business.