I don't think so. It's a laudable incremental advance in an area that computers (including artificial intelligence) are very good at: games with a constrained context and rules. By “constrained context” I mean that in order to play the game of Go (or most other games), you don't need to know anything beyond the rules of the game.
That's very different, for example, than holding a conversation with someone. In any randomly chosen human conversation, you'll find that in order to have that conversation the people involved must share an enormous amount of context. For example, if we were talking politics, you'd have to know the names of the politicians involved, something about their beliefs and pronouncements, what offices they were running for, who the voters were ... I could go on for a long time like that.
It's not that I think Google's accomplishment is unimpressive – it's very impressive, in the same way that being the first to walk on the moon is impressive. But ... it's less a fundamental advance and more an incremental improvement. I'd be much more impressed if someone rolled out a computer that could talk with me in a natural way on even some limited topic. As I've said before, I'm skeptical that such a thing will ever happen through digital computers as we know them today. Want some evidence for my skepticism? Spend a few minutes trying to ask some questions of Siri, Cortana, or Alexa (the voice enabled “digital assistants” from Apple, Microsoft, and Amazon, respectively). Are they helpful? Sometimes. Are they amusing? Most definitely. Do they seem like a human assistant? Hell, no. Not even close.
So ... congratulations, Google! But I'm still skeptical about the breadth of artificial intelligence's utility, and I still cringe when I hear people equate today's artificial intelligence with human intelligence. Those two concepts – today, at least – aren't in the same category...
No comments:
Post a Comment