> Okay and why exactly would anybody be interested in that? Sounds like a pointless hypothetical
"Okay, so this new attack vector that hasn't existed before will appear, but maybe no one will ever want to use it" isn't a great argument.
I'm not sure how to explain it better that total aggregation of all data that's ever existed on you into one convenient blob is dangerous. Even answering your direct question is trivial: imagine that you break up with someone and they use your public data to stalk you. Or you have a conflict with your relatives, who can then use that data to see where you might be and what you might be up to.
But the obvious further implications of systems like those are second-, third-, fourth-order processing. So, we live in a society where you can look people up by their photos. Up next, the videos they appeared in will be location-tagged, with other services putting two and two together and selling everyone's location data, just put a photo in and you get a convenient timeline. Other services may sell information on what people and groups you were seen associating with. These are all very convenient things to offer - now, police investigations go a lot easier, breaking up inconvenient groups is less troublesome, employers and landlords can screen you for 'undesirable' qualities before ever seeing you. Lots of possibilities, lots of people who would love to use that.
> someone puts up a service that let people upload a picture of your face and have it return a dossier with every bit of video you've ever appeared in
Okay and why exactly would anybody be interested in that? Sounds like a pointless hypothetical