How to Save Photojournalism
It might be time to dust off your film camera
I was at a San Francisco Giants game recently. As often happens to me when I watch a baseball game, my mind spent some time wandering, and at one point I found myself thinking that it would be really fun to see professional players stage an exhibition match using only aluminum bats. Of course, it would be a home run-fest, but to see what guys at that level could do with that kind of equipment might be fun for an afternoon.
Equipment regulations are a critical part of many sports, of course — no aluminum baseball bats in baseball, racquets must be a particular size in tennis, no brass knuckles in football, etc. But the other day, after watching another mid-blowing demo that made me fear for the death of photojournalism, I started thinking that maybe equipment regulations are a way to tend to some of the issues that AI is bringing to the veracity of image making. Specifically, if we want to save photojournalism maybe we need to require journalistic images to be shot on film and printed in a darkroom. Of course film prints can be manipulated, but only with great skill, and to a fairly limited degree by today’s standards.
“But!” you might say, “journalism is delivered digitally, so there are plenty of places in the pipeline to the end-user where deceptive changes could be made.” That’s true, but if publications offered high-res downloads of images, we could easily examine them ourselves. Film images differ greatly from digital images when you view them at high magnification because of film grain. Silver halide crystals clump together in very specific ways, and if that grain has been manipulated, it’s not hard to see evidence of alterations. Digital manipulations result in stretched and twisted film grain.
“Aha!” you might then say, “what about film printers?” That’s a good question. Hollywood proved long ago that you can create computer-generated images, print them to film, project that onto a wall and convince people of dinosaurs, galaxies far far away and much more. Someone could easily manipulate an image, or use AI to create something from scratch, print that to film to produce a negative, then print that negative to paper to produce an actual silver print. The only way to detect if it were fake would be to look for digital artifacts from the original image that were preserved through the printing process. In other words, we’d be back to the same forensics problem that we have now, but with the extra complication of film grain possibly obscuring digital artifacts.
But manipulation is not a new issue in photojournalism — it’s been there from the very beginning of photography, and so journalists have long had practices in place to try to counter it. Many of these are still legitimate and valid regardless of whether you’re shooting on digital, film, or making stuff up from scratch using an AI. For example, reputable journalistic agencies are skeptical of single images submitted by a photographer. They want to see an entire roll of film, or an entire sequence of numbered digital images. With a full batch of shots they can better understand the context of a situation to determine if the photographer’s choice of framing was representative of the reality of that scene.
By that standard, if you wanted to fake an image of a politician engaged in a Satanic ritual you wouldn’t be able to get by with producing a single perfect image, you would have to create dozens of perfect images, and the whole set would have to be a believable progression of compositions through a believable reality of a situation. That is suddenly a much more complicated problem. Insurmountable? Of course not, given enough time and skill. But it would filter out the casual “enthusiast” who wants to troll.
Even if someone does concoct that perfectly believable series of fake images — one that reflects a true process of how a photographer would capture a single hero shot — there are plenty of other checks that an editor can make. Does the photographer have any proof that they were at the location where they claim the photos were shot? Motel and restaurant receipts? Proof of travel? Any other shots from the area? Again, all of that could be faked or gathered, but it would take a lot of work.
Giving journalists the equipment requirement of film cameras would not solve the problem, but it would make manipulation much more complicated, and so would reduce the number of occurrences.
Of course the bigger problem with journalism in general — whether it’s written, photographed or videoed — is not its creation but the fact that consumers are not more skeptical. Until they take the time to fact check stories on their own, until they investigate who funds particular stories and what economic biases their journalistic sources might have, until they realize that a blogger or social media poster is not the same as an actual journalist — until those things become common practice on the part of the consumer, those people will still be susceptible to being tricked and fooled. All of that is a function of education, of course, and we don’t currently live in an age that offers egalitarian education so I’m not going to hold my breath for the consumer side of journalism to improve.
As player skill in professional sports has been improved and refined, the rules of some games have had to be altered to accommodate the increased ability of those players. You usually can’t introduce more performance to a system without having to modify the system to handle that increased performance. AI is a huge change in imaging performance and our current journalism system will definitely buckle under the strain of what it can do. The solution lies not in trying to prevent the use of AI — that’s impossible — but in altering the system to accommodate this new technology in a healthy way.
Whether that will happen remains to be seen, but if you’re a photojournalist, it might be time to dust off your film camera. You’ll have something to offer a potential publisher: a level of credibility that digital shooters might not have for long.
