How will AI affect our life?

There’s no evidence of AI getting out of control just because it becomes more powerful. The impact of AI will continue to be positive.

AlphaGo is more powerful at Go than Deep Blue was powerful at chess. But AlphaGo focuses its power on a predefined useful task. We don’t know what move AlphaGo is going to make, but we do know that it is going to play Go, which is a harmless activity.

All this talk about AI getting out of control is just speculation. People like to ask if AI were to get out of control, then what would happen? This question is a means for them to examine human nature. AI is a mirror that people hold up to see themselves in.

Some people think that intelligence is unpredictable by its nature. There is no evidence of such unpredictable behavior in programmable machines. So there is no more than a remote chance that AI could get out of control.

AI hasn’t shown itself to be more unpredictable than other kinds of programming. We will continue to make AI safe and predictable just as we always have with other kinds of software.

It is absurd to speculate about out-of-control AI just because AlphaGo won a board game. Yet we see talk like that in the news all the time. Reporters seriously ask whether the world is about to end just because of some trivial advance in programming. I wonder why they don’t treat other fields that way. It really makes no sense.

I do think that AI will surpass human intelligence within the next twenty years or so. But it will do so one step at a time, while remaining safe. Progress in AI has been safe in the past, and it should remain safe in the future.