It seems like a lot of the focus has moved from "find this exact image" to "find similar images," like you give it a picture of a Husky and it'll find more pictures of Huskies.
It seems like a lot of the focus has moved from "find this exact image" to "find similar images," like you give it a picture of a Husky and it'll find more pictures of Huskies.
I think maybe that's part of the picture.
Edit: I looked up the details again. It was an FSB officer named Andrei Burlaka, involved in the shooting down of MH17. Someone in Russia were selling their TV set online, and had taken a picture of the TV for the ad. It just so happened that Burlaka was on TV exactly that moment. Yandex images indexed the ad, leading the investigators to identify him.
Let me give you an example - you have a photo of someone. Should Google give you their facebook profile in the results if you search for the picture? Their AI is certainly capable of this, but some people would be understandably creeped out by it.
To be clear what I'm talking about, with tineye.com you can feed it an image and TinEye will show you all places on the internet that no only use that image - also use parts/portion of that image.
I think what you're describing is just straight up image search.
It used to be much better than it is now though, I don't really bother with it these days.