Name: What Google's DeepMind 'Go' Loss Says About AI
Uploaded: 2016-03-14
Duration: 2 min 54 s
Description: Fortune’s Robert Hackett and Andrew Nusca debate.

[ MUSIC ] [ SOUND ] A human beat a computer in Google's latest deep mind artificial intelligence challenge. Robert, is AI not ready for prime time? I don't think AI is ready for prime time, Andrew. Come on. Now, I mean, let's be clear. The computer still took home the trophy. It sure did. It beat the world champion in a majority of these matches with the ancient Chinese board game, Go. However, we're building these machines that are really good at doing specific, narrow tasks. We have machines that can whoop you at checkers, machines that can beat you at chess, and machines that utterly trounce you in Jeopardy. But in order for A. I. to really reach the next level, I think we need systems that are gonna be as adaptive and intelligent and Imaginative as humans can be. I agree with you in theory but I fundamentally disagree with the core of your argument which, you seem to suggest that AI can't be useful if it's not operating at a human level. Look, AI's all around us right now. If you check your Facebook newsfeed, the items that are displayed there are dictated by AI. If you use Siri and ask it to send a note to a loved one it is using AI to parse the words in your speech. It's all around us. Yes, these systems are simplistic, compared to the theoretical maximum. They are rudimentary, but they are learning. That is what distinguishes AI From the rest. Let's take a look at history though. I mean right now AI hype is at a fever pitch but we've been here before back in the 70s. There was a lot of hype around AI. People were saying machines are gonna be just as intelligent as humans just a few years from now. All of a sudden they didn't meet up with the expectations, the grand expectations that they had laid out and funding dried up and research stopped. This happened again in the 90s. We went through AI winters. I think that could be just around the corner again. But you're suggesting that we haven't really advanced enough for it to all have been worthwhile. Of course we have. I mean, the AI today is way better than all of the decades past that you mentioned. I mean, come on. Does Lebron James make every basket to be the best basketball player on the court? No. Does Watson, IBM's Watson, when it was playing Jeoprady in 2011, did it win all of the questions and get them all right to beat Ken Jennings? No. The fact of the matter is the trend line is in the right place, the systems are advancing, we are getting better, and business is paying attention. Here's where you and I differ though. If we are going to live in a world where we're letting artificially intelligent systems. Driver less cars and autonomous vehicles ferry us to and fro. Bring it on. They cannot get if one of these vehicles gets into a crash that leads to a human death that would be completely impermissible. That's true. We have to make sure that these AI systems are inarguably better than humans. That's a fair point. For one show that's more man in machine, come to fortune. com for more Tech Debate. [ MUSIC ]

What Google's DeepMind 'Go' Loss Says About AI

Fortune’s Robert Hackett and Andrew Nusca debate.