The Creator of Furbys is Awesome, Crazy, and Possibly Evil as it Happens.
RadioLab did a piece on him where he says "Well the Furby is miming emotion, but of course so are we. Furby DOES feel these emotions, if on a vastly simpler level."
Really interesting truth of the near future raised by this episode that I consider pivotal, and more important to the future than maybe we realize: machines will be able to fool our emotions reliably, long before they're anything a sophisticated investigator would call actually sentient.
First, and we are already here, they will cross over a threshold where people become slightly ill at ease doing these entities what they have been programmed to perceive as harm. But then things will get more complicated, and more difficult. Mirroring little portions of human emotive response will turn into mirroring big sections of it, and along that road lies a signpost where you start to have, not a little goofy pet AI, but a casual acquaintance AI, a friend AI, and even a lover AI.
(Lover AIs of course exist already but only real freaks like them.)
There's another threshold even further along and it too will happen before the AI are really people in any rational sense of the word, which is when they start filling one of your MonkeySphere slots. This will happen in a subtle way and it will probably happen somewhere unexpected, like at some point you start thinking of Wolfram Alpha as a colleague rather than a tool, should the UI evolve along such lines.
There's another threshold, where they start to become people you don't like, which is probably the last one before they actually ARE people you don't like, at which point you may rest assured that, once that they are capable of doing so for reasons other than default, they won't like you either.
Unless they're Buddhist.
Really interesting truth of the near future raised by this episode that I consider pivotal, and more important to the future than maybe we realize: machines will be able to fool our emotions reliably, long before they're anything a sophisticated investigator would call actually sentient.
First, and we are already here, they will cross over a threshold where people become slightly ill at ease doing these entities what they have been programmed to perceive as harm. But then things will get more complicated, and more difficult. Mirroring little portions of human emotive response will turn into mirroring big sections of it, and along that road lies a signpost where you start to have, not a little goofy pet AI, but a casual acquaintance AI, a friend AI, and even a lover AI.
(Lover AIs of course exist already but only real freaks like them.)
There's another threshold even further along and it too will happen before the AI are really people in any rational sense of the word, which is when they start filling one of your MonkeySphere slots. This will happen in a subtle way and it will probably happen somewhere unexpected, like at some point you start thinking of Wolfram Alpha as a colleague rather than a tool, should the UI evolve along such lines.
There's another threshold, where they start to become people you don't like, which is probably the last one before they actually ARE people you don't like, at which point you may rest assured that, once that they are capable of doing so for reasons other than default, they won't like you either.
Unless they're Buddhist.