If there was a time to plead to your friends to delete the photo of you passed out in your car in a wet Halloween costume or the ones where you were going through that unibrow phase, now would be that day. A new program created by Parham Aarabi at the University of Toronto has an algorithm that can sift through billions of photos on social media sites and find photos of you, even if they’re not tagged. Yes, even if they are outstandingly heinous and taken at a really bad angle, you will never be able to hide them.
The seemingly omniscient program utilizes “tag locations to quantify relationships between individuals.” What this technological hooby-whats-it does is track the relationships between those people who you’re already in tagged photos with. Here’s a fictional example: you’re in over 50 photos with Sarah and always in close proximity, then in about 80 photos with Jason and always right by one another, and then you appear in a photo with both Sarah and Jason apple picking. If only Sarah and you are tagged in the orchard photo, this program will still be able to pick out Jason in the photo. Because this computer program knows how important Jason is to you and is betting on the fact that you were in, yet again, another photo with him. This isn’t facial recognition software either, it’s a relational social image search that uses information about how we interact with each other to create more efficient search results.
So will this really nifty yet terrifying search engine behave like socialized medicine and remain in Canada and far, far away from unflattering photos you don’t even realize are uploaded on social media? Nope! Aarabi’s next plan is to sell the algorithm to large image databases like Facebook and Flickr for easier, quick photo searches for friends, employers, and potential dates alike. Now would be the time to carefully reflect on your freshman year album entitled, “Beer Boogie.” Consider yourself warned.
Image via Flickr.