Artificial intelligence has just evolved in a huge way. The NY Times reports that scientists at Google and at Google Stanford University have developed a software that can teach itself to recognize what’s happening in photographs. Google’s old-school image-recognition software could only teach itself to identify individual objects. (After seeing 10 million images from YouTube, it taught itself to recognize cats. Ha!) This new program is braniac in comparison; in most cases, the software’s photo captionswere pretty much the same as those of human observers. It’s very cool, but maybe a little unsettling? I just wonder, when the software describes “A pizza sitting on top of a pan on top of a stove,” what does that mean to the software? What is its understanding of what pizza, a pan, and a stove are?
I guess one day very soon, I’ll be able to ask it myself.