In the late 1990s I was part of a group that was thinking about how cameras might be able to sign images so that later editing could be detected. The motive then was connected to using images as evidence. But I guess you could also use it to mark edited images in some way.
Would apps like Instagram be "improved" if retouched images displayed a prominent "FAKE" icon of some sort?
What does "fake" mean? Even if you're taking as-camera images (which have already had de-Bayer and some level of dynamic range and denoise processing, which may be considerable and involve machine learning!), then you can't tell what manipulation has been applied to the scene. You can rent little "sets" for taking photography in. Props and clothes can be borrowed. And so on.
Authenticity is a product, just like any other. And it's increasingly hard to say what it even means in the 21st century.
But I always use McDonald's (although any restaurant can be used here) as the core of the conundrum of fake vs. real. Look at a corporate shot of a cheeseburger and compare it to reality. The marketing cheeseburger is the platonic ideal of the product -- it's fresh, perfectly prepared with care with all of the ingredients visible. The cheeseburger on the ground looks like a crumpled napkin in comparison.
Is the marketing cheeseburger fake?
When I look at some of the more image-conscious Facebook/Instagram people, some of them are kinda insane with editing and retouching. My wife has an old acquaintance who produces a highly "optimized" picture at least twice a month. (With substantial retouch to her face and usually other body "enhancements". In the age of Instagram+Photoshop, when does someone cross the line from marketing cheeseburger to fraud?
Go further down the rabbit hole on that one; is makeup a form of fraud? Heels / lifts? Push up bras? Instagram has simply accelerated and magnified what already happens in good ol' regular life (and introduced mass distribution of it).
This reminds me of Magritte's The Treachery of Images [1] and Calvin and Hobbe's Photo Lies [2]. All images are inherently lies - colors aren't completely accurately reproduced, perspectives are changed - so what is the threshold for manipulating the subject until the photographed subject should be deemed "fake?"
I think that it lies on a spectrum, that a dolled up cheeseburger could be more fake in the sense that a telephoto black-and-white picture could be less authentic than a color photo taken at a focal length of ~40mm, but it depends on the viewer whether to accept or reject a photo as true.
So many late nights at that McDonalds, and in that area. (Dundas and Bathurst - Toronto). Even this video is an idealization. That corner gets grimy at night. At least it used to around that time. And that burger she bought was not what they look like coming out of there. I have to laugh a little...
At least they don't use waxes in the model burgers anymore. That's a step toward honesty, at least.
When my kids asked me why we didn't have a real Christmas tree this year, I told them, "we do have a real Christmas tree. It's made of real plastic and wire."
To me this is like saying, "I'd like TV shows to all be shot live with no editing, and I'd like my books and news articles to be first drafts." I mean there's a place for those things (live theater or SNL, and poetry slams where stuff is improvised), but I certainly wouldn't want the majority of my media to be like that. Editing is more than just removing the ugly and emphasizing the beautiful. It's a way to clarify what point you're making.
No. Would restaurant food be improved if all meals came with "FAKE" sticker, because they do not match the raw ingredients?
Even straight out of camera image has been processed _a lot_.
I don't really care "how it looked in reality", because what is reality? All that matters is what the photographer wanted to show.
But the world is heavily influenced by all those fake images. People may know something is fake, yet being exposed to it again and again do change their expectation and behaviour.
It's like ad. You think it doesn't work. That you are not that stupid. But eventually, the brute force wins.
Both Canon and Nikon offered expensive cryptographic addons for their pro cameras that signed out-of-camera images. It never really caught on due to cost and, more significantly, the acceptance that some postprod as always required ( except for straight-to-press sports shots ).
for younger people, this is already solved by noticing each online system corruption of the image. e.g. some social networks add a greenish tint, other have a bug that adds an also greenish (but stronger) tiny to only the last column of pixels, etc. then by looking at those artifacts you can know not only if it's a repost but for how long it's been reposted.
it's a skill we, older, not so immersed, totally miss.
Would apps like Instagram be "improved" if retouched images displayed a prominent "FAKE" icon of some sort?