I remember posting pretty much this in the Reg's AlphaGo article
And being voted down by a lot of ninnies who don't get the difference between being programmed to play chess or play Go or play Jeopardy, and being programmed to LEARN, and being able to learn to do all those things plus many more. Because AlphaGo was programmed to "learn" from past games, apparently that meant it was an AI. :P
We are decades away from a true AI.