The AI Debate

With technology such as Siri and miniature unmanned areal vehicles with self-adjusting capabilities on the rise, a fairly difficult question is becoming more and more prevalent. While all of these programs are undoubtedly smart and they do what they where meant to do well, but do they count as AI? For that matter, what would be true AI?

TED is trying to answer this, by giving a challenge to engineers worldwide. To build and program a robot capable of delivering a TED talk. This project is still in it's infancy, and TED is still using polls to set the rules for what constitutes, in the public's mind, as AI. After all, you can always program something to give  pre-written speech, but that wouldn't be considered AI. So just giving a speech and walking off stage, in my mind at least, does't cut it.

What would be cool is to program a robot to preform in a debate. Instead of regurgitating information, the program would have to adapt and counter an actual, live human being in a verbal spar. It doesn't have to win, even humans have to loose debates, but it should at least put up enough adaptability to be an actual threat to it's opponent's point of view.

But even that has limits. That AI would be good for it's intended purpose, but what then? You couldn't stick the robot in, say, a rescue mission. That's not what it's built for. So, what are we supposed to do?

Well, the smartest and most complex thing we humans know about is our own brain. If we where to scan our brains, every minute detail, every neuron and synapse, and where to translate that into binary, and run THAT as a program, then that, in my opinion, would be as complex as we could conceivably make AI.