By Micah Sébastien Zhang
JRNL 200 Blog #09
Based on Option 1.
This article first appeared on VROOM’s UBC Blog on November 5, 2024.
I was a bit doomed as I saw numerous ultra-realistic AI-generated images, thinking that this might be the end of journalism today.
Seeing countless examples of people mistaking AI-generated or deepfake images in real life made me realize that my career would be way more treacherous ahead. If that’s as hard as cramming my 1500-word essay 3 hours before my deadline (yep I’ve done that before in my Year 1; it is bad kids), then it would be like climbing Mount Everest singlehandedly when those pictures started to appear in my family’s group chat with my grandpas and grandmas believing them.
Those ideas were overriding my mind…until one idea slided in.
Being an amateur photographer, I’ve gotten around the technical lingos. Having enough vocabulary dexterity around words like shutter, aperture, exposure, ISO has made my through, completing numerous contributions to my schools’ events (which means hearing "WE LOVE YOUR PICTURES SOOOO MUCH!!!" but ended up getting no credits). After all these years getting around with digital photography and cameras, one functionality got into my attention, as I believe it might be a solution to this entire comtemporary counterfeit media madness.
And that is (drum rolls pls): raw images files.
Raw files are basically the purest form of images produced from cameras. We see images normally in conventional and universal formats like jpeg or png; raw files come in lots of variations (depending on each camera’s manufactuer). But all raw files come with one key characteristic: their ability of not being altered.
Raw files are essentially containing all the original photographic information on a camera’s sensor. It benefits a lot to the post-production of photos, yet it also determined the fact that once you decided to alter images, you can’t save them into raw files format again. A blog post on DPReview citing an article from National Geographic (THERE IS PAYWALL AHHHH) suggested that it is National Geographic’s way to determine an image’s authenticity: photographers need to hand in photos’ original raw files if there are disputes over their authenticities. Since all raw files are all produced from a digital source (hence, a camera), it wouldn’t be that hard to apply the same technology on generative AIs, right?
Right?
Now get to work C2PA I’m putting all my hopes on you.
VROOM 