Algorithmic Lookalikes

I’m back visiting my old home town of Nashville for a few days, and had a super-fun breakfast conversation with Mary Laura Philpott this morning during which she mentioned that she sometimes gets automatically tagged in photos as Nashville mayor Megan Barry. The two bear a passing resemblance but not enough that you’d probably think to comment on it. She also mentioned that she has occasionally heard that she looks like Reese Witherspoon (although she never hears that in Nashville — Witherspoon’s home town).

I also used to hear all the time that I looked like Sandra Bullock, and the joke used to go that I could be her security double — the person who goes out the front door to throw off the fans and paparazzi so the actual star can sneak out the back door. (I’ve always wished I could be mistaken for Connie Britton just for the hair, but I don’t have the patience for hot rollers.)

Anyway, it occurred to me as Mary Laura and I were chatting that there’s a new kind of double: the facial recognition algorithm double. Facial recognition algorithms have become a routine part of our social media and personal photo library management, but they’re going to show up more and more in varied aspects of our lives, from surveillance to shopping. And the idea that you can “pass for” someone else — and that someone else could pass for you — is a tad troubling, isn’t it?

After all, there’s not much we can do about it, unless we have reconstructive surgery and hot-roll our hair and even then we might start getting tagged as Jocelyn Wildenstein or something, so we should probably just accept whatever Doppelgänger fate hands us and get on with life. The machines don’t know whether we’re the mayor of Nashville or the star of “Nashville” or just visiting Nashville.