I enjoyed your discussion. I was a journalist for 30+ years, the first seven of which I also was a photojournalist. The ethical questions you pose were real even back in the mid- to late 1980s when I stepped into my career. I always had difficulty accepting environmental portraits, for example, that were set up. They were planned to create an effect, and I always thought they should be labeled photo illustrations simply because of that. Later, as an editor, I often surprised photographers with the types of questions I asked, specifically because of some of the concerns you both discussed here. There is, I think, a vast difference between editing a photo for clarity -- to reveal detail that might be obscured in shadow, for example, which I would argue to absolutely acceptable -- and editing to create an effect or mood. The latter, in my opinion, is when the photographer steps out of the realm of photojournalism and into the realm of illustration. Both can be tools in journalism, but I think illustrations need to be clearly labeled as such in order to be honest with the readers/viewers. Sine elements of AI can be honest photographic tools when it is used with integrity. I'm thinking specifically of AI software than can sharpen an image, or remove noise. But when it is used to manufacture something that is not real, then it should be identified clearly as an illustration. Thanks again for posting this discussion! You both rock!
The first thing taught in my History of Photography class at my Uni is never to trust any photo. We saw hundreds of examples of doctored and stagged photos during WW1 and WW2, even some from the Civil War.
As a landscape photographer I'm seeing the trend on social media that the general public assumes the fantastic AI images with double rainbows, lightning bolts, fantastic shaped clouds are real. I'm amazed how many thousands of folks praise the photographer for the great work...
I personally think that if you are adding something to the photo that wasn't there at the moment you took the photo (or if you missed the item in the photo and add it later with AI), you need to disclose. I'm so sick of looking at these photos online and wondering if they are real or not. I want people to disclose if the photo was altered beyond basic editing and cropping - mainly meaning items added but also items removed.
One of the biggest obstacles is trying to identify when a "significant" edit is made, because that implies judgment. An edit that is significant to one person, in one context, may be completely irrelevant in another. Even something as basic and otherwise acceptable as cropping could exclude extremely relevant information in photojournalism. It's all about the context.
@@parkereayup. But you could make the argument you could have taken the same image without cropping (reframing, zooming, etc). But of course what about changing Colors? Aren’t those considered removing something and adding something new? What about masked editing?
I determine it based upon the type of photo. I have artistic/abstract photos where I will blow out the saturation to where it no longer looks the same, but haven't added anything that wasn't there. If I do a landscape I try to make it look as much like I remember it as possible, as I feel like blowing out a sunset or sunrise sky and comping that over more naturally lit mountains is deceptive unless you somehow indicate you did it.
@@michaelt1103 for sure. I was thinking of wildlife/landscape that are intended to represent actual wildlife and locations. I think it's great to be creative with photos. So yes, I think it does depend on the type of photo and intention.
When I was in school for photography I had to videotape all the taking of my images because the teachers thought they were photoshopped. Eventually they realized that everything I did was not photoshopped I just like to take unusual shots and did unusual images, but I was also really good at using Photoshop
Decades ago the father of Polaroid, Edmund Land, remarked that if a photograph wasn't a Polariod, there was no way to be sure it hadn't been altered. Photo manipulation, retouching, compositing...all used to take lots of skill before pixel level editing became possible.
I felt the same way when Photoshop first became a thing. Even shooting film I always tried to choose the best emulsion to get the results I wanted. Photography will always be an interpretation of reality. My beef is adding elements that were not captured in the original photo. That is NOT a photograph, it's a photo based unreality.
Another famous instance of staging was the nature film by Disney about lemmings. For decades people thought when lemmings get overpopulated a few run off a cliff and the rest just follow. Being a "lemming" became known as someone who blindly follows the crowd. There's even a series of video games about this. But it came out later that the filmmakers were chasing the lemmings off the cliff in order to get a good shot. Lemmings DO have booms in population but they don't jump off a cliff, groups migrate to a new area and inevitably a few fall into the ocean and die but it's nothing like what was portrayed and what entered the public mind.
"When I post a black and white photo, do I need to provide a disclaimer that the scene was actually in color? If I choose a wide aperture, should I disclose that the background wasn't actually out of focus? 🤣
This is an interesting conversation for photographers to have. Two things are important, in my view. A) a photograph is an instant in time of a visual scene with boundaries. That applies to any photography made with any device. the luminance and color data collected in digital form is then used by software, either in the camera (JPG) or via editing software externally (Photoshop et al) to create a representation of that visual instant in time, with boundaries.) The ONLY person who knows the accuracy of the representation is the photographer that pressed the shutter release. All the rest of us are relying on their statements to determine its visual validity. Which leads to B. When a photographer is making an image to document the bounded instant in time they have legal and ethical responsibilities. When the photographer is making an image as the beginning point of a creative expression they have no legal or ethical obligations beyond their creative intent. For example, when photographing a house for a realtor, there are rules and laws that govern what is acceptable as a reasonable representation of the property at the time it's being sold. On the other hand a photographer making a picture of the same property for creative purposes is not constrained by those rules and laws. As a young commercial photographer I was hired by an attorney to make pictures of the site and environment in a store where their client had been injured. Even though the pictures were made on film and then printed on paper, I was required to appear in court, with the negatives and prints to testify that I had made the pictures and they were an accurate representation of the location. I expect something similar has to be done today, even with digital cameras. At the end of the day, it's the photographer determining real or contrived. AI is just another tool available to interpret our creative intent.
I worked hi-rise construction when I was younger, and I did often eat sitting on I-Beams or with my legs dangling off the side. First day on the job I climbed up 24 stories without a safety or tether. The i-beam photo may be fake, but we really did the stuff just like that.
Great Discussion! I think it gets complicated quickly and on many different levels. 1) If I correct geometric distortion from a wide angle lens image am I "faking" it or making it more real? The same can be said for chromatic aberration, high ISO noise, poor metering choice, etc., etc. 2) Tony touched on this, and I concur that often I'll make adjustments to an image to make it closer to what the scene ACTUALLY looked like than what the camera captured. 3) I'm pretty sure that painters who were commissioned to produce portraits of their clients 500 years ago, were taking liberties to flatter their benefactors. It was even know back then as A.I. (Artists Income).
In my view it’s pretty simple, correcting lens aberrations, especially distortion and CA, using considered faking anything. Whereas adding a composite element that wasn’t there is faking it. There is a grey area in the middle but the extremes are pretty obvious , to me at least.
I was commenting on the comedic angst they were showing. I recently lost a friend of 35 years to cancer, and just thought a reminder that this may not be as critical as some may feel would be out of place, offensive or boring. I stand corrected.@@ulyssesnathanialowen3831
I'm a retired Navy Photographer, retired 23 years ago. I got into Digital in 94, and immediately we saw it abused. As photographers, we had service record entries noting we would NOT manipulate imagery, as our PRIMARY job was documentation. Even in portraits, we weren't supposed to alter or remove blemishes, as even official portraits were regarded as legal documentation. In fact, Marine promotion board images that appeared to be manipulated would be disqualified. But when we did work that was purely artistic, we frequently had to disclose we had made manipulations. (I much preferred Aldus Photostyler to early Photoshop, though!!!) That said, nobody ever really thought much about darkroom manipulations, dodging, burnings, exposure correction color correction... I honestly can I don't miss the smell of the chemistry, but kind miss the magic of watching the print appear in the soup... All that said, one of the first lessons at Navy Photoschool included how much photography can lie, how much can you do to an image before nobody trusts it anymore...
These issues are very complex and difficult to sort out. Just the other day and turned my nose up at someone who put an obvious fake background behind some flying ducks in a social media post. During this podcast, I thought of the times I have done some serious editing on a animal photo to make the background less distracting. Using the remove tool in Photoshop, and a few other techniques, one can really change the entire feeling/mood in a simple bird photo. Now I feel bad for how I thought about what the other photographer did. Fortunately for me, I didn't go as far as to actually make a negative comment in response to the other photographer...I just moved on. The picture the other photography posted was actually more enjoyable than just a friggin blue sky would have been, so that is a plus, and it was actually the same motivation I had for changing my background. Do you really need to disclose this information for a simple social media post? I'm not making any money, I don't think the other photographer was making any money, and our viewers had something enjoyable to look at as they got through their days. Yet, in the moment, part of me felt like the other person was somehow cheating, while I didn't bother to look at myself in the mirror. Not good.
Interesting discussion and I don't really have any answers for the most part. I do think that the Samsung situation is a pretty clear cut line though, they misrepresented the capability of their camera, because it wasn't telling us it can create sharp moon pictures, it was telling us it can zoom in to a distance as far as the moon and get sharp pictures. It implied you could use this functionality on any zoomed image.
I've been watching a guy on UA-cam who goes into deep analysis of musical performances and the use of auto tuning and pitch correction and it's very interesting to see how closely this discussion parallels that debate. It's surprising how many world renowned vocalists use these technologies either by choice or forced by the studios. He even disclosed the use of these methods during broadcasts of talent programs!
His name is Fil, but he works under "Wings of Pegasus". His work is very relevant to this discussion. He's covered many recording artists. He electronically isolates out their vocals and shows the waveforms of the notes they're hitting. His studies of Billy Joel at the Grammys, Karen Carpenter's live and studio vocals and Roy Orbison singing Walk On are good examples of how AI'ing performances diminishes them.
There's also a very long but very good article in the Feb 5, 2024 "New Yorker" magazine on how the executive level of the music business is coping with AI called: "The Next Scene - Lucian Grainge helped the music industry survive file-sharing. Now he wants to do the same with A.I."
I In the first few minutes, where you talk about Samsung's premise that photos are all 'fake', I wanted to add another perspective ... Think about images taken by space telescopes (Hubble, JWST, etc). My understanding is that these images are composed of layers, each taken in monochrome but using specific filters, then each is assigned a color, and finally composited together. The results are certainly beautiful, but nothing the human eye would ever see. Do we questions whether they're real? Not really.
I pretty clearly draw the line by adding any objects into an image that were not there, while removing things like a small part of a window leaking into the image or removing a pencil in the corner of the floor is fine for me. Mostly like you'd clean up someone's face in a portrait to just show how a scene would have ideally looked. This way, I don't feel like making something up that wasn't there, but only removing something annoying that was there by extending something that was there. If you remove a piece of trash from the floor by just replacing it with more of the existing floor, it's just different from putting a whole new teddy bear on that floor.
There's perhaps play no way to capture the totality of a person's life with a single frame but that's very different from the idea of intentionally spreading misinformation or propaganda. When you're creating an image of people working on a skyscraper in order to bolster the national spirit in the wake of a construction boom or when you're trying to make someone feel a certain way about poverty by portraying it with fake images you can do real damage. It only flashed on the screen for a moment but apparently the Afghan girl was jailed and everyone in the West only new false information about her. National Geographic may not have told us the everything about her or the war but they certainly could have done a better job if they were intentionally lying and ruined her life. That's the difference between whether or not there's a real picture or not. Real pictures don't have to be 100% a reflection of the totality of the universe that the subject inhabits, but they cannot contain intentional lying meant to spread false narratives or harm people.
Agree 100% People should also view pictures with "some" skepticism and assume there might be some perspective trickery going on at least. Regarding "The Beam" Having a construction backbround, I wasn't as "overawed" as some . . . I just thought it was a cool photo.
I took a photo 8 years ago of a seaplane driving down the road outside of my house. Everyone trolls especially went wild accusing me of Photoshop. It was straight from the camera, one end of the road is a Marina and other end is the airfield. Nothing photoshopped or AI but it is amazing how quickly people troll unique photos
When I was a darkroom / analog photographer, I used to show the black outer edge of my photo to “flex” that I composed my shot in the camera, not in the darkroom. Now, when I compose the digital photo as I take it, there’s no way for me to show it to viewers the way I used to show the black darkroom frame.
My thoughts: Comptetitions have their own rules, so I won't dwell on that. Generally we just don't want to feel tricked. But, our brain works very differently from how a camera works. When you're out watching stuff, your brain has to process a lot of info, and some of it just isn't that important from moment to moment and so your brain discards it. In a landscape there could be a soda can in a bush that you don't notice at all among the myriad of impressions and experiences of viewing a scene. But capture it in a photo, and it's there forever as part of the landscape. It will get noticed eventually. And, sure, that can really was there, and to edit it out is to change how the landscape looked at that precise moment and angle. However, if you walk for 20 miles through nature and there is only one can, a single photo showing that can would misrepresent the rest of that landscape for what it was. We simply observe things differently in real life compared to in a photo. The photo is, in my mind, supposed to recreate, in a fleeting moment, a generalised impression of reality as we live it. This is the case especially when it comes to how we see other people. Our brain is hardwired to hyperfocus on faces, because faces are important to us. So when we see a person in real life our brain focuses on changes in tone, mannerisms, expressions etc. and filters out unimportant things. We would struggle in social situations if our brain tried to take in all information available. Whereas the camera captures everything in one instance that can be scrutinised forever. Every blemish, stray hair, etc. When we LOOK at a person, we hardly ever notice these things unless we make a conscious effort to do so. So, to replicate how a someone looks to us in person, ironically, editing away the distractions makes sense. Or let me put it this way: There won't ALWAYS be a stray hair on the head of a person you know. There won't ALWAYS be some fluff on their shirt. You see them in poor light, in good light, freshly shaven and scruffy, with make-up and without, etc. You know what they look like despite these changes. The visage underlying the daily changes. A single photo can't get a "fair" impression of a person as we see them day-to-day. So, we remove the peculiarities of that moment.
Exactly. The argument about our faces is also why often they feel so weird when captured. You freeze a split second's tiny grimace that completely differs from what you actually notice and memorize with your brain. Photography has been a proxy of reality since it was invented.
I think the bigger problem is that every photo you take (or are about to take) is checked by Samsung to determine you are taking a photo/have taken a photo of the moon... Rather than the moon being fake.
The most vivid distinction would be if all photographs would include the raw file. So you could see the edit version and if you want you can see the original as well and compare them. Then you dont really have to manually explain all edits if you dont want to. The viewer can just see it. I am aware that there would be programs that would allow to edit the raw file as well so both would be fake but at least there would be something for the masses.
There is always going to be AT LEAST demosaicing done otherwise all our photos would be green. Programs like lightroom do that silently in the background but more advanced raw image processing software lets you turn it off or change the algorithm used. Because of this i tell people all the time there is no such thing as an unedited photo.
"There is no real picture, full stop." There is no real cell phone picture, full stop. Fixed it for you. Cell phone cameras spindle, fold, and mutilate images to the point where FOR CELL PHONE PICTURES, there ARE no real pictures. Which is one of the many reasons I'm not a fan of cell phone cameras.
I have taken multiple shots and merged them together many times. I do this to make a narrow image wider. Having said that, on 2 seperate occasions I merged photos of flying birds. One was a flock of geese that became 3 flocks of geese. On another occasion one buzzard became 4 buzzards. I have absolutely no idea how those pictures turned out like that. I'd love to pat myself on the back for my ingenuity, but both sets of pictures were total accidents.
The reason we have to edit photos to make them appear like what we saw ourselves is because even the best camera only has 12-15 stops of dynamic range, while the human eye has 21. In terms of math, it differs depending on the source. The issue is how subjective even the explanation of this can become. Having said that, what we see is so much different than how the camera sees. So I find in my photography trying to bring what I remember seeing back to the surface. Or, I could say, I try to bring out the image that's already there, to terribly rephrase a quote by Michelangelo.
I used to take group photos of100+ company employees. I'd always take multiple exposures so that I could drop in "better" facial expressions over terrible ones. I'll routinely delete things that I feel deter from a shot - like we used to do spotting a negative but these days you can remove more than spots. My most recent controversy with my brother involved a de-noised film shot where I removed substantial graininess.
Duade Paton does a straightforward job of 'disclosing' the reality of his photos in his UA-cam videos. First, he typically has a camera documenting himself in the field on his photo shoots. This is the equivalent of Tony's body camera. Second, he shows a final edited photo and then shows the original raw photo slowly blending into the crop and adjusted final photo.
Well, I believe you answered your own question at 2:41 into the video. It is in fact true that everything we experience is a sensory representation of the world, and the sensory projections go both ways into and out of our brains. We spend a great deal of time learning and confirming these representations. What's disturbing is the thought that someone or something might manipulate the sensory representation to provide a better equivalency to the real thing, one close enough to be the real thing. Alas, the great fear of AI!! It may corrupt the very nature of trust in our sensors.
As I watch more and more YT videos on photography and photo editing I can’t help but think how far photographers are getting from actual photography. Many of the comments say something like how this ot that person’s photo editing techniques has made them better photographers. REALLY? The photographer starts with a basic photo that is not very special. They often say how it’s not very good and lacks this or that quality and they then proceed to edit the photo to the point of NOT being closer to what he/she ACTUALLY WITNESSED and captured but to some extreme difference that was digitally altered to be what the photographer wished they had captured under those perfect circumstances. How is that significantly different from AI? One is faster and the other is done more slowly with post processing. Editing photos to way beyond making them look more faithful to what your eye/brain witnessed doesn’t make anyone a better photographer. It’s just making more people settle on being average photographers and editing their photos to their imagination’s content to make it look like they’re better photographers than they are. How is that good for the craft? There ARE still some very good photographers that put in the work necessary and come back to a scene time and again in order to actually capture a brief, real moment where everything came together to be captured and preserved in a photo or series of photos. Any skilled editor can fake a moment in time that never actually existed and pretend they "photographed" it.
FTR I super edit my photos to create a mood, not the actual scene and currently I don't composite. But I've learned a lot about the limitations of the programs through editing. Figuring out that I need to step to the left, knowing how to not over/underexpose in-camera, knowing how to get a composition I want with the only lens I could bring, embracing the onsite limitations. I'm not saying you're wrong, I'm being THAT person who says it's not that black and white.
The famous pictures of the woman with the two children and the girl with the green eyes are only 'untruthful' because they were represented as journalism. The distinction here is that between historical truth and something that might be called cosmic truth. If those two pictures were presented to the world without captions as simply slices of reality that invite the viewer to create the meaning then they become cosmic truth, which is where all great art lives. The 20th century writer Ernest Hemingway often talked about writing fiction that was "truer than if it had actually happened." Did the story he created in his anti-war novel "A Farewell To Arms" actually happen? No, but it was never meant to be historical truth, it was meant to be cosmic truth. Did Jesus actually wither the fig tree in the gospel of Mark? No, but the author of Mark (whose name wasn't Mark) wasn't writing history. It's true as allegory -- a type of cosmic truth. All of this is why photography has always had trouble being accepted as real art. Real artistic expression is unconcerned with historical truth, it's about cosmic truth. But because photography shows us what's really there and is a tool of journalism it relentlessy drags us away from the cosmic and toward the historical. I know -- too long, no one read it.
Point of interest- the older Hasselblad cameras in 120/220 format had two notches in the film aperture, which could be seen beyond the edge of the frame in any image that was contact printed or printed including the edge information (film type, frame number). Bronica used Hasselblad/Zeiss images in some of their ads, easy to tell due to the notches. But apparently Zenza-Bronica people did not realize it or thought no one else would. Of course if you had the negative or transparency in your hand, you could tell 100% of the time if it was a Hasselblad image. Anyway, in some ways this goes back further than digital.
Photos are an approximation of what we saw when we took a photo, not an accurate representation of what your eyes saw. Lenses capture aberrations, distortions, transmit color casts, and then go through a bayer sensor that have to go through a debayering process to make up data that wasn’t even there to begin with because it can’t capture all of the color information in one pixel. You then edit them in a RAW Processing software that will interpret and demosaic those files differently. By the time you share those images out on social media, it’s not “what you saw”. Shooting on film? The moment you invert a negative you have essentially edited it and it was also not what you captured. Slide film? Chromes have color casts too. Take it a step further if you print. What printer was used? Which inks, dye or pigment? How many inks, 4 or 12? What papers did you use, photo or “fine art”? So yes, photos technically aren’t real. Patrick wasn’t lying here. Although, I do agree that Samsung should have disclosed the fake moon popping up in people’s photos.
Use that quote to get out of any and all traffic violations caught on camera. Reminds of the cassette tape commercial that asked "Is it live or is it Memorex".
Re: personal boundaries: when you talk about "adding a fake crow" to your portraits, are you talking about taking another photo you took of a crow and compositing, generating an AI crow, or grabbing a crow image off the internet? These are important distinctions to make. (If you're calling it "YOUR art", I would hope only the first option would be the case). Either way, this is certainly an important and necessary discussion. Thanks.
I feel like a lot of photographers have moved into Digital photographers. This is when the image doesn't pass the eye test, like what you saw when you took the pic. Even sports photographers have gone to the extreme by over saturating colors especially for outside games. They look more like EA Sports games instead of a game seen in person.
I mostly shoot landscapes, and have no problem removing an object (power lines, trash cans, even people) but I would not add content (e.g., sky swap) without disclosing it. I would consider using generative expand and think that should be disclosed but, as Tony mentioned, how do you do that without people thinking the whole image is generated?
Samsungs current RAW support in ProMode is definitely fake: if you take dng + jpg there is nothing to recover in highlights and shadows, while a 3rd party app produces a useable raw.
According to English Language & Usage: Picture: An individual painting, drawing, or other representation on a surface, of an object or objects; esp. such a representation as a work of art. (Now the prevailing sense.) Image: An optical appearance or counterpart of an object, such as is produced by rays of light either reflected as from a mirror, refracted as through a lens, or falling on a surface after passing through a small aperture. Such an appearance may also be a mere subjective impression on the sense of sight, as an after-image (q.v.), and the negative image or accidental image seen after looking intently at a bright-coloured object, and having a colour complementary to that of the object. An image produced by reflexion or refraction is called in Optics a real image when the rays from each point of the object actually meet at a point, a virtual image when they diverge as if from a point beyond the reflecting or refracting body.
Before photography, many artists used camera lucida technology to help with their drawings. So, the relationship between "cameras" and "faking it" goes back centuries.
Not even close. Color Negative needs converting, and the moment you do that you have essentially created an interpretation of what the film saw. The inversion process can be done in multiple ways, each inversion giving different color results, and a different interpretation of the film. B&W film? We see in color, so not what you saw either. Slide film? All chromes have color casts that are characteristic of the film stock used. For example, Blue Ektachrome, Purple Provia, Orange Velvia, take your pic, none of them real.
Almost all cellphone photos are “fake” by some definition. Cellphone cameras add fake bokeh to simulate the depth of field effects in traditional cameras. You can also (for example) simulate long exposures using multiple images. These are examples of AI in photography, yet seem acceptable. If faking the moon should be marked as “fake” then these images should also be marked as “fake.” In the end of the day, what is acceptable or not depended on the customer. With cell phones, people just want cool looking pictures to share, and 99% of the time that’s all they care about. Astrophysicists don’t use cell phones. If they did, and sold it as accurate, that would be bad. However, for most people fake moons are just fine. In the end of the day, there is no such thing as real or fake. What matters is if the picture tells the story the client expects.
Of course! Here's easy proof. Most first camera owners think they can get or eventually get photos of similar to their favorite photographers with just the camera. Later they find out it's a lot of magic that comes with before/after taking the photo. In my opinion, the closest to real is street/reporting photography with no edits.
There is no simple way of guaranteeing that a photo is what it claims to be. Software can be used on an edited image to reformat the file so that it looks as if you haven't even edited it. So for photo competitions it is no longer good enough to submit the raw file to prove you met the rules.
The concept that the "camera cannot lie" seems to date back to the 19th century, but of course photographers can. Perhaps we should consider buying polaroid cameras (and getting back to shaking the prints again)
Even the original film negatives that I was fortunate enough to see in person from the first days of photography when there was glass negatives instead of film, these images were still either manipulated in some fashion or areas etched out to erase or hide certain elements of the image. So even to the first true photographs there is no such thing as an absolute real factual image that is not a direct interpretation by the photographer in some fashion or another. Even Ansel Adam’s used the darkroom to manipulate the negative to draw the image he saw in his mind out of the negatives that he was able to expose. Because sometimes there is not enough time to capture the image on the negatives the way you want or that even if the exposure is perfect, the image is not what you either saw or have constructed in your mind’s eye to create to present as the final image. From a graduate of the Ringling School of Art and Design with a BFA in Photography and Digital Imaging.
The photo is real which doesn’t mean the subject is real. Just like a painter or sculptor creating something that doesn’t exist. Both create beauty and are meant to be looked at. Example..Peter Lik
I think adding something that wasn't there (fake moon) fake object is where you cross the line and the photo is a fake. Removing small distracting items using generative ai is the same thing we did for years it was just much more tedious. The core for me is that I am making the photo show what I experienced. I saw the beautiful bird I didn't even notice the bits of flotsum and jetsum on the surface of the water. Interesting I ran into the trend to disbelieve. Another local photographer and I were at the Thunderbirds practice for the Daytona 500 and she captured this amazing shot of a Thunderbird jet and a bald eagle! The Birds of Prey forum decried it as a photoshop fake till... she posted shots of the shot on her camera with the image info on the display and I posted my 3 frames from continuous shooting showing the eagle (mine was not a nice it was not in focus because I was wide open on the jet) showing the same scene different angle. People are so suspicious now of photos and soon video and audio. Samsung technology would insert a moon where none existed and it wasn't optimizing the image it was substituting a complete fabrication. Samsung lied full stop.
Okay - take a seat - your friendly neighbourhood philosopher is on hand (seriously, it’s my job). Putting aside Samsung's facile attempt to avoid responsibility for an attempt to deceive, it is far worse than you imagine. The implicit photographic benchmark for what is "real" is what we (think we) experience when we look at a scene. An image counts as real if it looks like what we experience. But the frequencies of light that ultimately are the source of the image are not *really* red, or blue or any colour at all - the colours we experience along with everything else in our experience is a creation of the brain - an “interpretation”, if you will, of reality. And there is not even a one-to-one correspondence of frequencies of light to colour experiences, because our brain relativizes our experience to its presumptive context. The same frequency can look one colour in one context and a completely different colour in another; different frequencies will look the same colour in different contexts. The moment we click the shutter we have distanced the image from what we experience as reality. When we compose an image we select which parts of reality to include and which to exclude. But reality does not come with rectangular edges. Then there is long exposure and black and white which is not even how we experience the world, and all the editing techniques we use. The ethical issues you talk about are all to do with deception rather than reality, and that is a function of the expectations of the audience and the intentions of the image creator. Roughly speaking, an image is deceptive if: a) the image is not a veridical representation of the scene of which it purports to be (understood in terms of what the audience would expect that they might have experienced had they been behind the camera when the shutter was pressed). b) the audience might reasonably believe that what they are viewing is a veridical representation of the purported scene, and c) the photographer presents the image expecting the audience to interpret it as a veridical representation of the purported scene, whilst believing it to not be such. I won't bother to unpack “veridical representation” as that would take a while, but hopefully this provides a useful framework for thinking about the issues you raise. The other issue you raise concerns not the deception of the image so much as deception relating the image to the “image-maker” (is the image “mine” if it is AI generated?) or the image to the assumed context of its capture (e.g. staging). Mutatis mutandis, variations of the tri-partite conditions outlined above can be applied to these other dimensions of deception .
I don't use photography to copy a scene verbatim, I use it to see the scene through my mind's eye. To me, photography is as much an expression of art as an oil painting.
The question of disclosure is both simple and complex. It's simple, in that the ethical conundrum we face isn't over the image-making process used. It's when we intentionally misrepresent an image as something it isn't that we enter an ethical minefield. If an image is a product of the photographic process with only basic processing tools used, I don't think any disclosure is required beyond calling it a photo. When is greater disclosure required? It depends. In journalism, professional standards of objectivity and accuracy demand that photos accompanying news stories be held to a high standard. The integrity of the news agency is at stake. If the audience learns that an image presented as a photo of an astronaut walking on the Moon was actually an AI-generated image, that agency risks losing all credibility. In photo contests, there are often rules limiting the kind of processing that can be done to entered images. If elements that were present when the photo was made are removed or if elements that were not present are added, that misrepresents the moment captured in the photo. Such entries are often not allowed or become disqualified when discovered. The exact same image, however, would not raise any ethical issues if displayed in an art gallery. Society gives artists tremendous latitude in the processes used to create their original works. A Chuck Close photorealistic pencil sketch crosses no ethical lines. It looks like a giant black & white photo from a distance but is revealed to be a sketch upon closer examination. There is no deception. Even found objects - e.g. urinals - can be presented as "art" in exhibits, and recognized and reviewed as such. Art, by definition, challenges the societal boundaries. My personal policy is to not use processing tools to remove content within the frame or add content to the frame. That simplifies things for me. Others make full use of all the available image-making tools to create original works. My recommendation is to disclose the process used or the type of image made. If it's a composite photo, a collage, or AI-generated imagery, say so. Both the image-maker and the image should be recognized for the work done and the resulting product. To have a work recognized as something it isn't does a disservice to the artist, their methods and to those who view and appreciate the work.
Expectations for photographic truth were abandoned long ago. The absolute nearest any imaging process came to Truth was Kodachrome, and we abandoned it for convenience and quick turnaround time. Kodachrome was strictly lab processed, which was out of control or the ability for the user to alter or manipulate. It was 3 black and white emulsions to which pigments were added during processing, making it archival. And due to that process it had a known, reliable color palette. In summary, if you saw a Kodachrome slide of a UFO with Bigfoot, Jackie Onasis and Elvis stepping off the ramp, you might want to start looking for that UFO. (Or a near perfect model in a studio somewhere)
I'm embracing the features of generative AI. At a recent shoot, I covered a beat-up table with a big swatch of fabric. In post, I saw just how wrinkled the fabric was. I drew around it with the lasso tool, and typed in "smooth table cloth". The first set of options had a great table cloth - but it changed the color and I liked it. Done. Mind you, I'm a hobbyist who loves photography. I do consider it my art form. In my portraiture work, I have used freq separation for years, without reservation. I guess I haven't used AI in a way that's caused me to question it ... yet. Good topic -- oh yea, BTW, this is the 2nd video (I think) that you've referenced Marques Brownlee. Please note, the pronounciation of his first name is "Mark-Ezz". It's not Mark-Eese. Cheers!
I like the concept of exposure. If it has a defined exposure and some adjustments to the signal, it’s “real”. But if you add or subtract (including erasing a power line or adding a moon), that’s an “embellished” image. If I do an HDR or a mosaic, I think that is “enhanced”, and usually the photographer will tell you it’s an HDR or mosaic. I think if AI is cropping or fixing exposure, no big deal. But if it is adding or subtracting content in the image, it’s “embellished”. I also think if AI is “borrowing” content from other images it finds on the internet, it is plagiarizing. If AI studies other photos and mimics, that is fine, just like for humans.
I only use Photoshop Elements along with Topaz DeNoise AI to do minor adjustments to my bird photography. Shooting at 30fps with my Canon R7 permits me to get enough photos so I don't have much editing to do with the best ones. I don't like to over edit. Thanks for another amazing video Chelsea & Tony. Cheers from Montreal, Canada :)
‘Fake’ is getting to be an overused term. For instance I can point my mirrorless camera at the moon and take a picture. With a 600mm lens, details of the moon show up like the Jura Mountain chain and named craters. If you didn’t know the name of the craters, or mountain chains, or even believe that camera technology can produce such a detailed picture 200k miles away, you will scream that the image is fake. And if you have influence, you will get others to believe that it’s fake also. Our own eyes can be likened to sensors, except we can’t snap physical detailed pictures of what we see. We burn the images in our mind and create a memory with them. So we’ve gotten to a point where ‘fake’ and ‘real’ are being questioned, and when that bald eagle picks up a slice of pizza and flies off, if you aren’t a world renowned photographer people immediately doubt you captured an authentic moment. I think the sentiment came out because Samsung was happy to take people’s money and have them telling me they could take the same shot of the moon with their smartphone that I could with my mirrorless equipment. It sold a shortcut to the process and people love shortcuts.
Deceiving the eyes of the viewer is photography (art?) from the beginning. Multiple exposures, multiple films through the enlarger on a single print, special films, and so much more. AI is not new, we now have the computer power and memory to make it fast and simple. I tend to be a purist in that I want my print to be what I saw (or at least close to that).
The question is what is perception? We see with our eyes and memory of similar events. When you look at a natural scene you see into shadows and the brain lighten them and reverse for highlights. Then all are combined that becomes the memory. This is how we see. Now a camera can do the same.
The story a photo tells or is used to tell, is separate from the question of whether or not an image is a photo. When Chomet said, "No photos are real," he was either being intentionally evasive or simply ignorant. An AI-generated Moon image is not a photo. Samsung intentionally deceived the public by including that image in the promotional materials as representing the photos a Samsung smartphone camera will make. Of course, when you consider that smartphones long ago abandoned photography for the realm of digital art, is this the conversation we should be having? Shouldn't the discussion be, what tools remain for those who want to do photography; who want to make images using the photographic process?
My line is pretty simple for me: I want my photos to be a representation of what I saw with my eyes. This of course is subjective to my personal experience, which includes emotions. Put another way, I want someone to look at my photo and know that with timing, work, etc they could also have that experience in that location in the actual word. This often means I am desaturating certain images, especially astro photography. if I do something that is out of the realm of what someone could really see and experience on the planet earth, then I disclose that and say this is digital art. For human subjects, I try to do what could be done in camera with lenses - even if I cannot afford a $4000 portrait lens. But I really don't do portraits except for vacation photos! :D In the end it's all about discloser for me. If you're turning a sunset photos red when it had little color, just say you're doing that. If you don't WANT to disclose what you're doing, that might be a sign that you probably shouldn't be doing it at all.
Often, this depends on the usage of the image. I think disclosure is important... there's no reason to hide how the photo was made. For journalism or documentary work, I believe in staying true to the truth of the image... only cropping, toning for the light, some light dodging and burning, nothing that changes the meaning and message, nothing that removes elements of the image that were there. But it's fine to remove artifacts like dust particles that were on your sensor or lens, that aren't part of what your eye would see. When it comes to portraits that aren't Journalistic in nature, then I think you have leeway to meet the needs of your client with retouching, difussion/soft focus, removing blemishes, etc. If it's obvious that a photo is faked (distorted by lenses or forced perspective), that's not journalism or documentation... it's usually comedic or humorous. Art gallery work can involve nearly any type of photography or editing. It just needs to be labeled as such.
That quote could be from Goebbels. It's the rhetoric of someone who wants to skew your sense of reality with wordplay. It actually made me feel kind of gross
I call my "basic edits" "technical corrections", which points to things that couldn't be solved by the camera and lens itself (like color temp & gamut, exposure, sharpness, lens distortion). Everything on top of that would be "creative edits" where I take any liberty I want, since a photo is a photo is a photo. I want to create good pictures, not truthful visual representations of photons present at the time of exposure. That's just data science. Realism in that sense has always been a hoax - except maybe for journalism, where it is an ideal to strive towards, but not what is actually the case. I like the watermarking initiative in the context of reporting. Great discussion - thanks!
I agree in general with you both. There is a difference between going in under a specific class of photography (landscape, birds/animals, portraits, street) and using the photo with a process to overcome its limitations, or indeed, to use the limitations of the process (as with alternative processes) to create art. Continuing on art photo, me and my wife were just visiting Liljevalchs spring exhibition in Stockholm, and I was surprised and a little dismayed to find the information "manipulated photo" or "digitally enhanced photo" on some of the photographies. To me, entirely misplaced and taking away from the experience. Of course there will be manipulation with an art photo - that is what we do to get where we want to go...
Wow! You two do such a comprehensive job of expounding on an important subject in photography that it really opens my eyes to the social and political impact photography has on people. I thought little of the fake photo I did in high school of putting my brother on top of a Coke can in 1980 doing a double exposure on film with a Nikon F2. But, ethics are questioned when altered photos are published in the media and millions of people are influenced. Well done. Thank you
I never never add anything to my photos (no generative stuff), I rarely remove stuff and only if it is a distraction and a small one at that. I don't work with any clients so I don't feel the need to make any significant adjustments that stray too far from what I saw when I took the photo. That said, my photography is entirely subjective, and even though I make less edits than the average photographer I am in no way trying to keep the scene objective. At the end of the day you choose what to point the camera at and when to trip the shutter, so in my opinion the resulting image is always going to be subjective.
What’s disingenuous on Samsung’s part is they don’t acknowledge that this argument’s final resolution is that there is no reason to have a camera - at all! Just have a generative AI device. Comping a moon image into a photo is no different than going on a photo safari and having your camera comp in a “better” image of a zebra or lion. Go to Iceland and your camera comps in a “better” image of the aurora. WHAT is the point of buying their faux-camera device at all? The discussion about the “reality” of photos is a red herring, that’s a false argument. No experienced photographer argues that their photos are a reproduction of the world. That they are real or truth. They ARE real in that they capture light info and interpret it in a manageable way. That’s why we have white balance and color calibration. Photos are REPRESENTATIVE of a scene. Some cameras do better than others at this task. Some have a look we like because they create a unique look, eg: pinhole cameras. The argument that auto focus invalidates the accuracy of your image is bullshit. If Samsung wants to incorporate AI as an option that interprets and ads synthetic imagery into your scene, just be straight forward about it and let the marketplace decide if we are interested.
I go back to Edward Weston, one, if not the greatest photographer and artist of all time; yes, times do change, I know what you're thinking, but the facts do not change, a lie is still a lie, the truth is still the truth. Yes, Edward Weston manipulated his negatives in the darkroom but made contact prints from 8x10 inch negatives. He also used the simplest equipment and techniques perfectly, as stated in the documentary The Photographer. He didn't have to fake anything. He didn't have to add anything or subtract anything. He was burning and dodging. He was the master photographer, the master craftsman. Why do we have to progress in one direction towards less and less skill and more and more machine and less human being? Why can't we evolve backwards towards more skill in camera and by editing less and less. Progress doesn't have to be linear it can be circular. Can we find our way back to where we were, or will we abandon all of our humanity in favor of an imagined perfection or utopia using computer intelligence? I hope that we can get back to the basics!
We worry about AI assuming the fully AI generated will get us better photos than human taken. Technically that’s possible but we don’t get impressed by just perfect technologies. It’s way more complicated than perfect framing, exposure, and lights.
It's a question of semantics. How much or how little does a sensor capture versus how much processing happens? It reminds me of how much many of us hated overprocessed HDR photos back in the day.
That’s an abysmal rabbit hole to jump into, actually. And UA-cam is an unsuitable platform for discussing it. In the early days of photography, chemicals were the first step of manipulation. Nowadays the first steps are the way the sensor is built, followed up by the a/d converter and the firmware of the camera used. Technically spoken, you don’t even need a sensor to create a picture. Then the post processing steps in, enhancing the manipulation even more. Have you ever answered yourself the question of why you press the shutter release button in the first place? I mean for every picture you take? Sadly the reason is not captured in the image, so nobody will ever know. It’s even more sad, that I’m not allowed to communicate via text. Only to exchange information. But that’s another story for another time.
I tried taking a picture of a round tortilla with the S22 Ultra, S23 Ultra and the S24 Ultra. The cameras (or processing) DO try to sharpen the features of the tortilla to make it look more "moon-like" but in no scenario could I get any of the cameras to plop in a fake image of the moon. That's not what's going on at all. Same thing with a pure white circle on a black background. Samsung is not inserting fake images of the moon. It is using "AI enhancements" to pull out features when it detects you are taking a moon pic.
I've never heard you both discuss your thoughts on Vivian Maier. As you probably know, she was the woman whose (truly amazing) pictures were discovered after she died. She was not famous for her ability until after her death. This brings up a topic that I am constantly thinking about and that is the effect of others seeing your work and how that effects how you take future pictures. I really love your show and I especially like the somewhat existential questions (like the ones discussed in this episode) that you present to the community. Thanks! - Bryant
Photograph=an image created by capturing light. AI images are something else. AI images might replace photographs in many cases, but they are not photographs. A lot of your discussion was about how much processing you can do of a photograph before disclosing. AI image were never photos at all. All the discussion about "optical illusion" type photographs and mis-representation are not addressing the Photograph vs AI Image issue at all. Maybe not your goal, but ...
I'm a hobbyist and I've been photographing for several years now but this subject goes beyond photography. IMHO is about perception of reality. English is my second language so bear with me as this is a very interesting and difficult topic. Here is what I think: if you or anyone else decide to touch the concept of reality, first and foremost it needs to be defined. So... what's reality and what should be considered as real? In my opinion, once we press that button in the camera and have an image taken, it is a depiction of reality, therefore, is that considered real or fiction since it is only a snapshot? Is what we see completely real? Let's not even talk about post image manipulation. First, is it time a metric that we could use as a reality validation? Is the image considered real or fake due to the fact that all images belong to the past? Is reality intertwined with time? One could take a landscape photo of a sky and the next second could have a cloud. Does that mean when you look at the photo that it is still relevant to reality.... after a few days? This is a very difficult subject. Tony and Chelsea.... I ABSOLUTELY love your channel. I think I found you guys since you started. Thank you so much!
This is an important conversation to have. We, as professional photographers, have a profound responsibility to the truth in a world that increasingly views it with apathy. Another great video, guys!
My second thought is that no version of any image is “real” as no two single people have the exact same quality vision and no one’s mind interprets the raw information from your senses the same way and therefore no two people on the planet ever see the exact same image ever, but that is no excuse for not simple disclosing the fact that the image is “enhanced” as all authentic creators should do, but again this is a corp and not an individual artist. So again, it’s the lie that is offensive and not the message.
I think that Samsung guy got it all backwards, capturing a picture is what makes it real. Your eyes can deceive you, your memory can fail you, but a captured image should be proof that whatever you captured did happen.
NEW YEAR, NEW YOU! Go to squarespace.com/Chelsea & save 10% off your first website or domain with code “Chelsea"
"All photographs are accurate, none of them the truth" ~ Richard Avedon
This sounds like someone getting caught stealing and saying "Well technically, the concept of private property isn't real."
Private property IS a crappy concept. Personal property is fine but private property should be abolished.
@@MikeHunt4hire You're right. I'll come by and take all "your" stuff soon....😂
@@MikeHunt4hire my address is.... Canada🤣
You just described communism.
My thoughts exactly
I enjoyed your discussion. I was a journalist for 30+ years, the first seven of which I also was a photojournalist. The ethical questions you pose were real even back in the mid- to late 1980s when I stepped into my career.
I always had difficulty accepting environmental portraits, for example, that were set up. They were planned to create an effect, and I always thought they should be labeled photo illustrations simply because of that.
Later, as an editor, I often surprised photographers with the types of questions I asked, specifically because of some of the concerns you both discussed here. There is, I think, a vast difference between editing a photo for clarity -- to reveal detail that might be obscured in shadow, for example, which I would argue to absolutely acceptable -- and editing to create an effect or mood. The latter, in my opinion, is when the photographer steps out of the realm of photojournalism and into the realm of illustration. Both can be tools in journalism, but I think illustrations need to be clearly labeled as such in order to be honest with the readers/viewers.
Sine elements of AI can be honest photographic tools when it is used with integrity. I'm thinking specifically of AI software than can sharpen an image, or remove noise. But when it is used to manufacture something that is not real, then it should be identified clearly as an illustration.
Thanks again for posting this discussion! You both rock!
Remember the whole kerfuffle about the Matt Mahurin photo/illustration of O.J. Simpson for Time magazine.
The first thing taught in my History of Photography class at my Uni is never to trust any photo. We saw hundreds of examples of doctored and stagged photos during WW1 and WW2, even some from the Civil War.
Yes, and the Soviets were masters of picture manipulation. One day you are in a picture within a group, next day you never existed.
As a landscape photographer I'm seeing the trend on social media that the general public assumes the fantastic AI images with double rainbows, lightning bolts, fantastic shaped clouds are real. I'm amazed how many thousands of folks praise the photographer for the great work...
I personally think that if you are adding something to the photo that wasn't there at the moment you took the photo (or if you missed the item in the photo and add it later with AI), you need to disclose. I'm so sick of looking at these photos online and wondering if they are real or not. I want people to disclose if the photo was altered beyond basic editing and cropping - mainly meaning items added but also items removed.
totally agree, I feel like I'm questioning more and more pics because they just look too over done.
One of the biggest obstacles is trying to identify when a "significant" edit is made, because that implies judgment. An edit that is significant to one person, in one context, may be completely irrelevant in another. Even something as basic and otherwise acceptable as cropping could exclude extremely relevant information in photojournalism. It's all about the context.
@@parkereayup. But you could make the argument you could have taken the same image without cropping (reframing, zooming, etc). But of course what about changing Colors? Aren’t those considered removing something and adding something new? What about masked editing?
I determine it based upon the type of photo. I have artistic/abstract photos where I will blow out the saturation to where it no longer looks the same, but haven't added anything that wasn't there. If I do a landscape I try to make it look as much like I remember it as possible, as I feel like blowing out a sunset or sunrise sky and comping that over more naturally lit mountains is deceptive unless you somehow indicate you did it.
@@michaelt1103 for sure. I was thinking of wildlife/landscape that are intended to represent actual wildlife and locations. I think it's great to be creative with photos. So yes, I think it does depend on the type of photo and intention.
When I was in school for photography I had to videotape all the taking of my images because the teachers thought they were photoshopped. Eventually they realized that everything I did was not photoshopped I just like to take unusual shots and did unusual images, but I was also really good at using Photoshop
Decades ago the father of Polaroid, Edmund Land, remarked that if a photograph wasn't a Polariod, there was no way to be sure it hadn't been altered.
Photo manipulation, retouching, compositing...all used to take lots of skill before pixel level editing became possible.
I felt the same way when Photoshop first became a thing. Even shooting film I always tried to choose the best emulsion to get the results I wanted. Photography will always be an interpretation of reality. My beef is adding elements that were not captured in the original photo. That is NOT a photograph, it's a photo based unreality.
Another famous instance of staging was the nature film by Disney about lemmings. For decades people thought when lemmings get overpopulated a few run off a cliff and the rest just follow. Being a "lemming" became known as someone who blindly follows the crowd. There's even a series of video games about this. But it came out later that the filmmakers were chasing the lemmings off the cliff in order to get a good shot. Lemmings DO have booms in population but they don't jump off a cliff, groups migrate to a new area and inevitably a few fall into the ocean and die but it's nothing like what was portrayed and what entered the public mind.
I've seen nature films debunked a few times. Most of the stuff we see is manipulated for entertainment purposes. It's crazy what they hide from us.
"When I post a black and white photo, do I need to provide a disclaimer that the scene was actually in color? If I choose a wide aperture, should I disclose that the background wasn't actually out of focus? 🤣
This is an interesting conversation for photographers to have. Two things are important, in my view. A) a photograph is an instant in time of a visual scene with boundaries. That applies to any photography made with any device. the luminance and color data collected in digital form is then used by software, either in the camera (JPG) or via editing software externally (Photoshop et al) to create a representation of that visual instant in time, with boundaries.) The ONLY person who knows the accuracy of the representation is the photographer that pressed the shutter release. All the rest of us are relying on their statements to determine its visual validity. Which leads to B. When a photographer is making an image to document the bounded instant in time they have legal and ethical responsibilities. When the photographer is making an image as the beginning point of a creative expression they have no legal or ethical obligations beyond their creative intent.
For example, when photographing a house for a realtor, there are rules and laws that govern what is acceptable as a reasonable representation of the property at the time it's being sold.
On the other hand a photographer making a picture of the same property for creative purposes is not constrained by those rules and laws.
As a young commercial photographer I was hired by an attorney to make pictures of the site and environment in a store where their client had been injured. Even though the pictures were made on film and then printed on paper, I was required to appear in court, with the negatives and prints to testify that I had made the pictures and they were an accurate representation of the location.
I expect something similar has to be done today, even with digital cameras.
At the end of the day, it's the photographer determining real or contrived. AI is just another tool available to interpret our creative intent.
I worked hi-rise construction when I was younger, and I did often eat sitting on I-Beams or with my legs dangling off the side. First day on the job I climbed up 24 stories without a safety or tether. The i-beam photo may be fake, but we really did the stuff just like that.
Great Discussion! I think it gets complicated quickly and on many different levels. 1) If I correct geometric distortion from a wide angle lens image am I "faking" it or making it more real? The same can be said for chromatic aberration, high ISO noise, poor metering choice, etc., etc. 2) Tony touched on this, and I concur that often I'll make adjustments to an image to make it closer to what the scene ACTUALLY looked like than what the camera captured. 3) I'm pretty sure that painters who were commissioned to produce portraits of their clients 500 years ago, were taking liberties to flatter their benefactors. It was even know back then as A.I. (Artists Income).
In my view it’s pretty simple, correcting lens aberrations, especially distortion and CA, using considered faking anything. Whereas adding a composite element that wasn’t there is faking it. There is a grey area in the middle but the extremes are pretty obvious , to me at least.
We should also be very grateful that we live lives that allow us the luxury of debates like this.
bored with woke nonsense
I was commenting on the comedic angst they were showing. I recently lost a friend of 35 years to cancer, and just thought a reminder that this may not be as critical as some may feel would be out of place, offensive or boring. I stand corrected.@@ulyssesnathanialowen3831
Love that!
Is this a photo? -No this is Patrick.
I'm a retired Navy Photographer, retired 23 years ago. I got into Digital in 94, and immediately we saw it abused. As photographers, we had service record entries noting we would NOT manipulate imagery, as our PRIMARY job was documentation. Even in portraits, we weren't supposed to alter or remove blemishes, as even official portraits were regarded as legal documentation. In fact, Marine promotion board images that appeared to be manipulated would be disqualified.
But when we did work that was purely artistic, we frequently had to disclose we had made manipulations. (I much preferred Aldus Photostyler to early Photoshop, though!!!)
That said, nobody ever really thought much about darkroom manipulations, dodging, burnings, exposure correction color correction... I honestly can I don't miss the smell of the chemistry, but kind miss the magic of watching the print appear in the soup...
All that said, one of the first lessons at Navy Photoschool included how much photography can lie, how much can you do to an image before nobody trusts it anymore...
That's a cool story. You've been at this pretty much since the beginning!
These issues are very complex and difficult to sort out. Just the other day and turned my nose up at someone who put an obvious fake background behind some flying ducks in a social media post. During this podcast, I thought of the times I have done some serious editing on a animal photo to make the background less distracting. Using the remove tool in Photoshop, and a few other techniques, one can really change the entire feeling/mood in a simple bird photo. Now I feel bad for how I thought about what the other photographer did. Fortunately for me, I didn't go as far as to actually make a negative comment in response to the other photographer...I just moved on. The picture the other photography posted was actually more enjoyable than just a friggin blue sky would have been, so that is a plus, and it was actually the same motivation I had for changing my background. Do you really need to disclose this information for a simple social media post? I'm not making any money, I don't think the other photographer was making any money, and our viewers had something enjoyable to look at as they got through their days. Yet, in the moment, part of me felt like the other person was somehow cheating, while I didn't bother to look at myself in the mirror. Not good.
Interesting discussion and I don't really have any answers for the most part. I do think that the Samsung situation is a pretty clear cut line though, they misrepresented the capability of their camera, because it wasn't telling us it can create sharp moon pictures, it was telling us it can zoom in to a distance as far as the moon and get sharp pictures. It implied you could use this functionality on any zoomed image.
I've been watching a guy on UA-cam who goes into deep analysis of musical performances and the use of auto tuning and pitch correction and it's very interesting to see how closely this discussion parallels that debate. It's surprising how many world renowned vocalists use these technologies either by choice or forced by the studios. He even disclosed the use of these methods during broadcasts of talent programs!
His name is Fil, but he works under "Wings of Pegasus". His work is very relevant to this discussion. He's covered many recording artists. He electronically isolates out their vocals and shows the waveforms of the notes they're hitting. His studies of Billy Joel at the Grammys, Karen Carpenter's live and studio vocals and Roy Orbison singing Walk On are good examples of how AI'ing performances diminishes them.
There's also a very long but very good article in the Feb 5, 2024 "New Yorker" magazine on how the executive level of the music business is coping with AI called: "The Next Scene - Lucian Grainge helped the music industry survive file-sharing. Now he wants to do the same with A.I."
I In the first few minutes, where you talk about Samsung's premise that photos are all 'fake', I wanted to add another perspective ... Think about images taken by space telescopes (Hubble, JWST, etc). My understanding is that these images are composed of layers, each taken in monochrome but using specific filters, then each is assigned a color, and finally composited together. The results are certainly beautiful, but nothing the human eye would ever see. Do we questions whether they're real? Not really.
they are not adding things that were not there other than color. Some of the color is because the actual "color" is not visible to the human eye.
I pretty clearly draw the line by adding any objects into an image that were not there, while removing things like a small part of a window leaking into the image or removing a pencil in the corner of the floor is fine for me. Mostly like you'd clean up someone's face in a portrait to just show how a scene would have ideally looked. This way, I don't feel like making something up that wasn't there, but only removing something annoying that was there by extending something that was there. If you remove a piece of trash from the floor by just replacing it with more of the existing floor, it's just different from putting a whole new teddy bear on that floor.
There's perhaps play no way to capture the totality of a person's life with a single frame but that's very different from the idea of intentionally spreading misinformation or propaganda. When you're creating an image of people working on a skyscraper in order to bolster the national spirit in the wake of a construction boom or when you're trying to make someone feel a certain way about poverty by portraying it with fake images you can do real damage. It only flashed on the screen for a moment but apparently the Afghan girl was jailed and everyone in the West only new false information about her. National Geographic may not have told us the everything about her or the war but they certainly could have done a better job if they were intentionally lying and ruined her life. That's the difference between whether or not there's a real picture or not. Real pictures don't have to be 100% a reflection of the totality of the universe that the subject inhabits, but they cannot contain intentional lying meant to spread false narratives or harm people.
Agree 100% People should also view pictures with "some" skepticism and assume there might be some perspective trickery going on at least. Regarding "The Beam" Having a construction backbround, I wasn't as "overawed" as some . . . I just thought it was a cool photo.
I took a photo 8 years ago of a seaplane driving down the road outside of my house. Everyone trolls especially went wild accusing me of Photoshop. It was straight from the camera, one end of the road is a Marina and other end is the airfield. Nothing photoshopped or AI but it is amazing how quickly people troll unique photos
When I was a darkroom / analog photographer, I used to show the black outer edge of my photo to “flex” that I composed my shot in the camera, not in the darkroom. Now, when I compose the digital photo as I take it, there’s no way for me to show it to viewers the way I used to show the black darkroom frame.
Pretty much every smartphone distorts faces to make them more "favorable". At the very least they remove imperfections on skin.
I don't like it
My thoughts:
Comptetitions have their own rules, so I won't dwell on that.
Generally we just don't want to feel tricked. But, our brain works very differently from how a camera works. When you're out watching stuff, your brain has to process a lot of info, and some of it just isn't that important from moment to moment and so your brain discards it. In a landscape there could be a soda can in a bush that you don't notice at all among the myriad of impressions and experiences of viewing a scene. But capture it in a photo, and it's there forever as part of the landscape. It will get noticed eventually. And, sure, that can really was there, and to edit it out is to change how the landscape looked at that precise moment and angle. However, if you walk for 20 miles through nature and there is only one can, a single photo showing that can would misrepresent the rest of that landscape for what it was. We simply observe things differently in real life compared to in a photo. The photo is, in my mind, supposed to recreate, in a fleeting moment, a generalised impression of reality as we live it.
This is the case especially when it comes to how we see other people. Our brain is hardwired to hyperfocus on faces, because faces are important to us. So when we see a person in real life our brain focuses on changes in tone, mannerisms, expressions etc. and filters out unimportant things. We would struggle in social situations if our brain tried to take in all information available. Whereas the camera captures everything in one instance that can be scrutinised forever. Every blemish, stray hair, etc. When we LOOK at a person, we hardly ever notice these things unless we make a conscious effort to do so. So, to replicate how a someone looks to us in person, ironically, editing away the distractions makes sense.
Or let me put it this way: There won't ALWAYS be a stray hair on the head of a person you know. There won't ALWAYS be some fluff on their shirt. You see them in poor light, in good light, freshly shaven and scruffy, with make-up and without, etc. You know what they look like despite these changes. The visage underlying the daily changes. A single photo can't get a "fair" impression of a person as we see them day-to-day. So, we remove the peculiarities of that moment.
Exactly. The argument about our faces is also why often they feel so weird when captured. You freeze a split second's tiny grimace that completely differs from what you actually notice and memorize with your brain. Photography has been a proxy of reality since it was invented.
Very well said.
I think the bigger problem is that every photo you take (or are about to take) is checked by Samsung to determine you are taking a photo/have taken a photo of the moon...
Rather than the moon being fake.
Thank you. Now I understand why Tarkovski sculpted the light with Polaroid. (He had a book "Sculpting with light")
"Any static image was yesterday, today is another day" (Lars Willsen) ... relax people, take a deep breath :)
The most vivid distinction would be if all photographs would include the raw file. So you could see the edit version and if you want you can see the original as well and compare them. Then you dont really have to manually explain all edits if you dont want to. The viewer can just see it. I am aware that there would be programs that would allow to edit the raw file as well so both would be fake but at least there would be something for the masses.
Instead of apologizing and making things better, the guy just gaslighted a lot of photographers. Such BS of a guy.
There is always going to be AT LEAST demosaicing done otherwise all our photos would be green. Programs like lightroom do that silently in the background but more advanced raw image processing software lets you turn it off or change the algorithm used. Because of this i tell people all the time there is no such thing as an unedited photo.
"There is no real picture, full stop."
There is no real cell phone picture, full stop.
Fixed it for you. Cell phone cameras spindle, fold, and mutilate images to the point where FOR CELL PHONE PICTURES, there ARE no real pictures. Which is one of the many reasons I'm not a fan of cell phone cameras.
I have taken multiple shots and merged them together many times. I do this to make a narrow image wider. Having said that, on 2 seperate occasions I merged photos of flying birds. One was a flock of geese that became 3 flocks of geese. On another occasion one buzzard became 4 buzzards. I have absolutely no idea how those pictures turned out like that. I'd love to pat myself on the back for my ingenuity, but both sets of pictures were total accidents.
The reason we have to edit photos to make them appear like what we saw ourselves is because even the best camera only has 12-15 stops of dynamic range, while the human eye has 21. In terms of math, it differs depending on the source. The issue is how subjective even the explanation of this can become. Having said that, what we see is so much different than how the camera sees. So I find in my photography trying to bring what I remember seeing back to the surface. Or, I could say, I try to bring out the image that's already there, to terribly rephrase a quote by Michelangelo.
I used to take group photos of100+ company employees. I'd always take multiple exposures so that I could drop in "better" facial expressions over terrible ones. I'll routinely delete things that I feel deter from a shot - like we used to do spotting a negative but these days you can remove more than spots. My most recent controversy with my brother involved a de-noised film shot where I removed substantial graininess.
Duade Paton does a straightforward job of 'disclosing' the reality of his photos in his UA-cam videos. First, he typically has a camera documenting himself in the field on his photo shoots. This is the equivalent of Tony's body camera. Second, he shows a final edited photo and then shows the original raw photo slowly blending into the crop and adjusted final photo.
That sounds perfect! Great context.
Well, I believe you answered your own question at 2:41 into the video. It is in fact true that everything we experience is a sensory representation of the world, and the sensory projections go both ways into and out of our brains. We spend a great deal of time learning and confirming these representations. What's disturbing is the thought that someone or something might manipulate the sensory representation to provide a better equivalency to the real thing, one close enough to be the real thing. Alas, the great fear of AI!! It may corrupt the very nature of trust in our sensors.
Since this started I have been asked no less than 10 times if a photo I took was fake. 😢
As I watch more and more YT videos on photography and photo editing I can’t help but think how far photographers are getting from actual photography. Many of the comments say something like how this ot that person’s photo editing techniques has made them better photographers. REALLY? The photographer starts with a basic photo that is not very special. They often say how it’s not very good and lacks this or that quality and they then proceed to edit the photo to the point of NOT being closer to what he/she ACTUALLY WITNESSED and captured but to some extreme difference that was digitally altered to be what the photographer wished they had captured under those perfect circumstances. How is that significantly different from AI? One is faster and the other is done more slowly with post processing.
Editing photos to way beyond making them look more faithful to what your eye/brain witnessed doesn’t make anyone a better photographer. It’s just making more people settle on being average photographers and editing their photos to their imagination’s content to make it look like they’re better photographers than they are. How is that good for the craft? There ARE still some very good photographers that put in the work necessary and come back to a scene time and again in order to actually capture a brief, real moment where everything came together to be captured and preserved in a photo or series of photos. Any skilled editor can fake a moment in time that never actually existed and pretend they "photographed" it.
FTR I super edit my photos to create a mood, not the actual scene and currently I don't composite.
But I've learned a lot about the limitations of the programs through editing. Figuring out that I need to step to the left, knowing how to not over/underexpose in-camera, knowing how to get a composition I want with the only lens I could bring, embracing the onsite limitations.
I'm not saying you're wrong, I'm being THAT person who says it's not that black and white.
The famous pictures of the woman with the two children and the girl with the green eyes are only 'untruthful' because they were represented as journalism. The distinction here is that between historical truth and something that might be called cosmic truth. If those two pictures were presented to the world without captions as simply slices of reality that invite the viewer to create the meaning then they become cosmic truth, which is where all great art lives. The 20th century writer Ernest Hemingway often talked about writing fiction that was "truer than if it had actually happened." Did the story he created in his anti-war novel "A Farewell To Arms" actually happen? No, but it was never meant to be historical truth, it was meant to be cosmic truth. Did Jesus actually wither the fig tree in the gospel of Mark? No, but the author of Mark (whose name wasn't Mark) wasn't writing history. It's true as allegory -- a type of cosmic truth. All of this is why photography has always had trouble being accepted as real art. Real artistic expression is unconcerned with historical truth, it's about cosmic truth. But because photography shows us what's really there and is a tool of journalism it relentlessy drags us away from the cosmic and toward the historical. I know -- too long, no one read it.
Point of interest- the older Hasselblad cameras in 120/220 format had two notches in the film aperture, which could be seen beyond the edge of the frame in any image that was contact printed or printed including the edge information (film type, frame number). Bronica used Hasselblad/Zeiss images in some of their ads, easy to tell due to the notches. But apparently Zenza-Bronica people did not realize it or thought no one else would. Of course if you had the negative or transparency in your hand, you could tell 100% of the time if it was a Hasselblad image. Anyway, in some ways this goes back further than digital.
Photos are an approximation of what we saw when we took a photo, not an accurate representation of what your eyes saw.
Lenses capture aberrations, distortions, transmit color casts, and then go through a bayer sensor that have to go through a debayering process to make up data that wasn’t even there to begin with because it can’t capture all of the color information in one pixel.
You then edit them in a RAW Processing software that will interpret and demosaic those files differently.
By the time you share those images out on social media, it’s not “what you saw”.
Shooting on film? The moment you invert a negative you have essentially edited it and it was also not what you captured. Slide film? Chromes have color casts too.
Take it a step further if you print. What printer was used? Which inks, dye or pigment? How many inks, 4 or 12? What papers did you use, photo or “fine art”?
So yes, photos technically aren’t real. Patrick wasn’t lying here.
Although, I do agree that Samsung should have disclosed the fake moon popping up in people’s photos.
Use that quote to get out of any and all traffic violations caught on camera. Reminds of the cassette tape commercial that asked "Is it live or is it Memorex".
Re: personal boundaries: when you talk about "adding a fake crow" to your portraits, are you talking about taking another photo you took of a crow and compositing, generating an AI crow, or grabbing a crow image off the internet? These are important distinctions to make. (If you're calling it "YOUR art", I would hope only the first option would be the case). Either way, this is certainly an important and necessary discussion. Thanks.
So you get a stock photo that you don’t own?
AI is not Photography. Full stop.
I feel like a lot of photographers have moved into Digital photographers. This is when the image doesn't pass the eye test, like what you saw when you took the pic. Even sports photographers have gone to the extreme by over saturating colors especially for outside games. They look more like EA Sports games instead of a game seen in person.
I mostly shoot landscapes, and have no problem removing an object (power lines, trash cans, even people) but I would not add content (e.g., sky swap) without disclosing it. I would consider using generative expand and think that should be disclosed but, as Tony mentioned, how do you do that without people thinking the whole image is generated?
Samsungs current RAW support in ProMode is definitely fake: if you take dng + jpg there is nothing to recover in highlights and shadows, while a 3rd party app produces a useable raw.
According to English Language & Usage:
Picture: An individual painting, drawing, or other representation on a surface, of an object or objects; esp. such a representation as a work of art. (Now the prevailing sense.)
Image: An optical appearance or counterpart of an object, such as is produced by rays of light either reflected as from a mirror, refracted as through a lens, or falling on a surface after passing through a small aperture. Such an appearance may also be a mere subjective impression on the sense of sight, as an after-image (q.v.), and the negative image or accidental image seen after looking intently at a bright-coloured object, and having a colour complementary to that of the object. An image produced by reflexion or refraction is called in Optics a real image when the rays from each point of the object actually meet at a point, a virtual image when they diverge as if from a point beyond the reflecting or refracting body.
Before photography, many artists used camera lucida technology to help with their drawings. So, the relationship between "cameras" and "faking it" goes back centuries.
Film gets the closest to something “real”
Not even close. Color Negative needs converting, and the moment you do that you have essentially created an interpretation of what the film saw. The inversion process can be done in multiple ways, each inversion giving different color results, and a different interpretation of the film.
B&W film? We see in color, so not what you saw either.
Slide film? All chromes have color casts that are characteristic of the film stock used. For example, Blue Ektachrome, Purple Provia, Orange Velvia, take your pic, none of them real.
Almost all cellphone photos are “fake” by some definition. Cellphone cameras add fake bokeh to simulate the depth of field effects in traditional cameras. You can also (for example) simulate long exposures using multiple images. These are examples of AI in photography, yet seem acceptable. If faking the moon should be marked as “fake” then these images should also be marked as “fake.” In the end of the day, what is acceptable or not depended on the customer. With cell phones, people just want cool looking pictures to share, and 99% of the time that’s all they care about. Astrophysicists don’t use cell phones. If they did, and sold it as accurate, that would be bad. However, for most people fake moons are just fine. In the end of the day, there is no such thing as real or fake. What matters is if the picture tells the story the client expects.
Of course! Here's easy proof. Most first camera owners think they can get or eventually get photos of similar to their favorite photographers with just the camera. Later they find out it's a lot of magic that comes with before/after taking the photo. In my opinion, the closest to real is street/reporting photography with no edits.
There is no simple way of guaranteeing that a photo is what it claims to be. Software can be used on an edited image to reformat the file so that it looks as if you haven't even edited it. So for photo competitions it is no longer good enough to submit the raw file to prove you met the rules.
The concept that the "camera cannot lie" seems to date back to the 19th century, but of course photographers can. Perhaps we should consider buying polaroid cameras (and getting back to shaking the prints again)
Even the original film negatives that I was fortunate enough to see in person from the first days of photography when there was glass negatives instead of film, these images were still either manipulated in some fashion or areas etched out to erase or hide certain elements of the image. So even to the first true photographs there is no such thing as an absolute real factual image that is not a direct interpretation by the photographer in some fashion or another. Even Ansel Adam’s used the darkroom to manipulate the negative to draw the image he saw in his mind out of the negatives that he was able to expose. Because sometimes there is not enough time to capture the image on the negatives the way you want or that even if the exposure is perfect, the image is not what you either saw or have constructed in your mind’s eye to create to present as the final image. From a graduate of the Ringling School of Art and Design with a BFA in Photography and Digital Imaging.
The photo is real which doesn’t mean the subject is real. Just like a painter or sculptor creating something that doesn’t exist. Both create beauty and are meant to be looked at. Example..Peter Lik
A lot of us are capturing memories they are not fake.
I think adding something that wasn't there (fake moon) fake object is where you cross the line and the photo is a fake. Removing small distracting items using generative ai is the same thing we did for years it was just much more tedious. The core for me is that I am making the photo show what I experienced. I saw the beautiful bird I didn't even notice the bits of flotsum and jetsum on the surface of the water. Interesting I ran into the trend to disbelieve. Another local photographer and I were at the Thunderbirds practice for the Daytona 500 and she captured this amazing shot of a Thunderbird jet and a bald eagle! The Birds of Prey forum decried it as a photoshop fake till... she posted shots of the shot on her camera with the image info on the display and I posted my 3 frames from continuous shooting showing the eagle (mine was not a nice it was not in focus because I was wide open on the jet) showing the same scene different angle. People are so suspicious now of photos and soon video and audio. Samsung technology would insert a moon where none existed and it wasn't optimizing the image it was substituting a complete fabrication. Samsung lied full stop.
Okay - take a seat - your friendly neighbourhood philosopher is on hand (seriously, it’s my job). Putting aside Samsung's facile attempt to avoid responsibility for an attempt to deceive, it is far worse than you imagine. The implicit photographic benchmark for what is "real" is what we (think we) experience when we look at a scene. An image counts as real if it looks like what we experience. But the frequencies of light that ultimately are the source of the image are not *really* red, or blue or any colour at all - the colours we experience along with everything else in our experience is a creation of the brain - an “interpretation”, if you will, of reality. And there is not even a one-to-one correspondence of frequencies of light to colour experiences, because our brain relativizes our experience to its presumptive context. The same frequency can look one colour in one context and a completely different colour in another; different frequencies will look the same colour in different contexts.
The moment we click the shutter we have distanced the image from what we experience as reality. When we compose an image we select which parts of reality to include and which to exclude. But reality does not come with rectangular edges. Then there is long exposure and black and white which is not even how we experience the world, and all the editing techniques we use.
The ethical issues you talk about are all to do with deception rather than reality, and that is a function of the expectations of the audience and the intentions of the image creator. Roughly speaking, an image is deceptive if:
a) the image is not a veridical representation of the scene of which it purports to be (understood in terms of what the audience would expect that they might have experienced had they been behind the camera when the shutter was pressed).
b) the audience might reasonably believe that what they are viewing is a veridical representation of the purported scene, and
c) the photographer presents the image expecting the audience to interpret it as a veridical representation of the purported scene, whilst believing it to not be such.
I won't bother to unpack “veridical representation” as that would take a while, but hopefully this provides a useful framework for thinking about the issues you raise.
The other issue you raise concerns not the deception of the image so much as deception relating the image to the “image-maker” (is the image “mine” if it is AI generated?) or the image to the assumed context of its capture (e.g. staging). Mutatis mutandis, variations of the tri-partite conditions outlined above can be applied to these other dimensions of deception .
I don't use photography to copy a scene verbatim, I use it to see the scene through my mind's eye. To me, photography is as much an expression of art as an oil painting.
The question of disclosure is both simple and complex. It's simple, in that the ethical conundrum we face isn't over the image-making process used. It's when we intentionally misrepresent an image as something it isn't that we enter an ethical minefield. If an image is a product of the photographic process with only basic processing tools used, I don't think any disclosure is required beyond calling it a photo. When is greater disclosure required? It depends.
In journalism, professional standards of objectivity and accuracy demand that photos accompanying news stories be held to a high standard. The integrity of the news agency is at stake. If the audience learns that an image presented as a photo of an astronaut walking on the Moon was actually an AI-generated image, that agency risks losing all credibility.
In photo contests, there are often rules limiting the kind of processing that can be done to entered images. If elements that were present when the photo was made are removed or if elements that were not present are added, that misrepresents the moment captured in the photo. Such entries are often not allowed or become disqualified when discovered.
The exact same image, however, would not raise any ethical issues if displayed in an art gallery. Society gives artists tremendous latitude in the processes used to create their original works. A Chuck Close photorealistic pencil sketch crosses no ethical lines. It looks like a giant black & white photo from a distance but is revealed to be a sketch upon closer examination. There is no deception. Even found objects - e.g. urinals - can be presented as "art" in exhibits, and recognized and reviewed as such. Art, by definition, challenges the societal boundaries.
My personal policy is to not use processing tools to remove content within the frame or add content to the frame. That simplifies things for me.
Others make full use of all the available image-making tools to create original works. My recommendation is to disclose the process used or the type of image made. If it's a composite photo, a collage, or AI-generated imagery, say so. Both the image-maker and the image should be recognized for the work done and the resulting product. To have a work recognized as something it isn't does a disservice to the artist, their methods and to those who view and appreciate the work.
Expectations for photographic truth were abandoned long ago. The absolute nearest any imaging process came to Truth was Kodachrome, and we abandoned it for convenience and quick turnaround time. Kodachrome was strictly lab processed, which was out of control or the ability for the user to alter or manipulate. It was 3 black and white emulsions to which pigments were added during processing, making it archival. And due to that process it had a known, reliable color palette. In summary, if you saw a Kodachrome slide of a UFO with Bigfoot, Jackie Onasis and Elvis stepping off the ramp, you might want to start looking for that UFO. (Or a near perfect model in a studio somewhere)
I'm embracing the features of generative AI. At a recent shoot, I covered a beat-up table with a big swatch of fabric. In post, I saw just how wrinkled the fabric was. I drew around it with the lasso tool, and typed in "smooth table cloth". The first set of options had a great table cloth - but it changed the color and I liked it. Done. Mind you, I'm a hobbyist who loves photography. I do consider it my art form. In my portraiture work, I have used freq separation for years, without reservation. I guess I haven't used AI in a way that's caused me to question it ... yet. Good topic -- oh yea, BTW, this is the 2nd video (I think) that you've referenced Marques Brownlee. Please note, the pronounciation of his first name is "Mark-Ezz". It's not Mark-Eese. Cheers!
I like the concept of exposure. If it has a defined exposure and some adjustments to the signal, it’s “real”. But if you add or subtract (including erasing a power line or adding a moon), that’s an “embellished” image. If I do an HDR or a mosaic, I think that is “enhanced”, and usually the photographer will tell you it’s an HDR or mosaic. I think if AI is cropping or fixing exposure, no big deal. But if it is adding or subtracting content in the image, it’s “embellished”. I also think if AI is “borrowing” content from other images it finds on the internet, it is plagiarizing. If AI studies other photos and mimics, that is fine, just like for humans.
I only use Photoshop Elements along with Topaz DeNoise AI to do minor adjustments to my bird photography. Shooting at 30fps with my Canon R7 permits me to get enough photos so I don't have much editing to do with the best ones. I don't like to over edit. Thanks for another amazing video Chelsea & Tony. Cheers from Montreal, Canada :)
‘Fake’ is getting to be an overused term. For instance I can point my mirrorless camera at the moon and take a picture. With a 600mm lens, details of the moon show up like the Jura Mountain chain and named craters. If you didn’t know the name of the craters, or mountain chains, or even believe that camera technology can produce such a detailed picture 200k miles away, you will scream that the image is fake. And if you have influence, you will get others to believe that it’s fake also.
Our own eyes can be likened to sensors, except we can’t snap physical detailed pictures of what we see. We burn the images in our mind and create a memory with them.
So we’ve gotten to a point where ‘fake’ and ‘real’ are being questioned, and when that bald eagle picks up a slice of pizza and flies off, if you aren’t a world renowned photographer people immediately doubt you captured an authentic moment.
I think the sentiment came out because Samsung was happy to take people’s money and have them telling me they could take the same shot of the moon with their smartphone that I could with my mirrorless equipment. It sold a shortcut to the process and people love shortcuts.
Deceiving the eyes of the viewer is photography (art?) from the beginning. Multiple exposures, multiple films through the enlarger on a single print, special films, and so much more. AI is not new, we now have the computer power and memory to make it fast and simple. I tend to be a purist in that I want my print to be what I saw (or at least close to that).
The question is what is perception? We see with our eyes and memory of similar events. When you look at a natural scene you see into shadows and the brain lighten them and reverse for highlights. Then all are combined that becomes the memory. This is how we see. Now a camera can do the same.
The story a photo tells or is used to tell, is separate from the question of whether or not an image is a photo. When Chomet said, "No photos are real," he was either being intentionally evasive or simply ignorant. An AI-generated Moon image is not a photo. Samsung intentionally deceived the public by including that image in the promotional materials as representing the photos a Samsung smartphone camera will make. Of course, when you consider that smartphones long ago abandoned photography for the realm of digital art, is this the conversation we should be having? Shouldn't the discussion be, what tools remain for those who want to do photography; who want to make images using the photographic process?
My line is pretty simple for me: I want my photos to be a representation of what I saw with my eyes. This of course is subjective to my personal experience, which includes emotions. Put another way, I want someone to look at my photo and know that with timing, work, etc they could also have that experience in that location in the actual word. This often means I am desaturating certain images, especially astro photography. if I do something that is out of the realm of what someone could really see and experience on the planet earth, then I disclose that and say this is digital art.
For human subjects, I try to do what could be done in camera with lenses - even if I cannot afford a $4000 portrait lens. But I really don't do portraits except for vacation photos! :D In the end it's all about discloser for me. If you're turning a sunset photos red when it had little color, just say you're doing that. If you don't WANT to disclose what you're doing, that might be a sign that you probably shouldn't be doing it at all.
Often, this depends on the usage of the image. I think disclosure is important... there's no reason to hide how the photo was made.
For journalism or documentary work, I believe in staying true to the truth of the image... only cropping, toning for the light, some light dodging and burning, nothing that changes the meaning and message, nothing that removes elements of the image that were there. But it's fine to remove artifacts like dust particles that were on your sensor or lens, that aren't part of what your eye would see.
When it comes to portraits that aren't Journalistic in nature, then I think you have leeway to meet the needs of your client with retouching, difussion/soft focus, removing blemishes, etc.
If it's obvious that a photo is faked (distorted by lenses or forced perspective), that's not journalism or documentation... it's usually comedic or humorous.
Art gallery work can involve nearly any type of photography or editing. It just needs to be labeled as such.
That quote could be from Goebbels. It's the rhetoric of someone who wants to skew your sense of reality with wordplay. It actually made me feel kind of gross
I call my "basic edits" "technical corrections", which points to things that couldn't be solved by the camera and lens itself (like color temp & gamut, exposure, sharpness, lens distortion). Everything on top of that would be "creative edits" where I take any liberty I want, since a photo is a photo is a photo. I want to create good pictures, not truthful visual representations of photons present at the time of exposure. That's just data science. Realism in that sense has always been a hoax - except maybe for journalism, where it is an ideal to strive towards, but not what is actually the case. I like the watermarking initiative in the context of reporting. Great discussion - thanks!
"How much editing did you do on this image?" "Just enough to crash my laptop"
I agree in general with you both. There is a difference between going in under a specific class of photography (landscape, birds/animals, portraits, street) and using the photo with a process to overcome its limitations, or indeed, to use the limitations of the process (as with alternative processes) to create art. Continuing on art photo, me and my wife were just visiting Liljevalchs spring exhibition in Stockholm, and I was surprised and a little dismayed to find the information "manipulated photo" or "digitally enhanced photo" on some of the photographies. To me, entirely misplaced and taking away from the experience. Of course there will be manipulation with an art photo - that is what we do to get where we want to go...
I think this video is vital on calling out all the fake advertisement. There are probably so many lies to uncover. Thank you!
I want AI to turn my wife into Margo Robbie every time I take a double selfie 😂
Very sad
Wow! You two do such a comprehensive job of expounding on an important subject in photography that it really opens my eyes to the social and political impact photography has on people. I thought little of the fake photo I did in high school of putting my brother on top of a Coke can in 1980 doing a double exposure on film with a Nikon F2. But, ethics are questioned when altered photos are published in the media and millions of people are influenced. Well done. Thank you
Thank you!
Samsung took a big hit of the devil’s lettuce before answering that question.
I never never add anything to my photos (no generative stuff), I rarely remove stuff and only if it is a distraction and a small one at that. I don't work with any clients so I don't feel the need to make any significant adjustments that stray too far from what I saw when I took the photo. That said, my photography is entirely subjective, and even though I make less edits than the average photographer I am in no way trying to keep the scene objective. At the end of the day you choose what to point the camera at and when to trip the shutter, so in my opinion the resulting image is always going to be subjective.
What’s disingenuous on Samsung’s part is they don’t acknowledge that this argument’s final resolution is that there is no reason to have a camera - at all! Just have a generative AI device.
Comping a moon image into a photo is no different than going on a photo safari and having your camera comp in a “better” image of a zebra or lion. Go to Iceland and your camera comps in a “better” image of the aurora. WHAT is the point of buying their faux-camera device at all?
The discussion about the “reality” of photos is a red herring, that’s a false argument. No experienced photographer argues that their photos are a reproduction of the world. That they are real or truth.
They ARE real in that they capture light info and interpret it in a manageable way. That’s why we have white balance and color calibration. Photos are REPRESENTATIVE of a scene. Some cameras do better than others at this task. Some have a look we like because they create a unique look, eg: pinhole cameras.
The argument that auto focus invalidates the accuracy of your image is bullshit.
If Samsung wants to incorporate AI as an option that interprets and ads synthetic imagery into your scene, just be straight forward about it and let the marketplace decide if we are interested.
Waves hand "That's not the photo you are looking for".
"How can photographs be real if our eyes aren't real."
I go back to Edward Weston, one, if not the greatest photographer and artist of all time; yes, times do change, I know what you're thinking, but the facts do not change, a lie is still a lie, the truth is still the truth. Yes, Edward Weston manipulated his negatives in the darkroom but made contact prints from 8x10 inch negatives. He also used the simplest equipment and techniques perfectly, as stated in the documentary The Photographer. He didn't have to fake anything. He didn't have to add anything or subtract anything. He was burning and dodging. He was the master photographer, the master craftsman. Why do we have to progress in one direction towards less and less skill and more and more machine and less human being? Why can't we evolve backwards towards more skill in camera and by editing less and less. Progress doesn't have to be linear it can be circular. Can we find our way back to where we were, or will we abandon all of our humanity in favor of an imagined perfection or utopia using computer intelligence? I hope that we can get back to the basics!
We worry about AI assuming the fully AI generated will get us better photos than human taken. Technically that’s possible but we don’t get impressed by just perfect technologies. It’s way more complicated than perfect framing, exposure, and lights.
It's a question of semantics. How much or how little does a sensor capture versus how much processing happens? It reminds me of how much many of us hated overprocessed HDR photos back in the day.
That’s an abysmal rabbit hole to jump into, actually. And UA-cam is an unsuitable platform for discussing it.
In the early days of photography, chemicals were the first step of manipulation.
Nowadays the first steps are the way the sensor is built, followed up by the a/d converter and the firmware of the camera used.
Technically spoken, you don’t even need a sensor to create a picture.
Then the post processing steps in, enhancing the manipulation even more.
Have you ever answered yourself the question of why you press the shutter release button in the first place?
I mean for every picture you take?
Sadly the reason is not captured in the image, so nobody will ever know.
It’s even more sad, that I’m not allowed to communicate via text. Only to exchange information.
But that’s another story for another time.
I tried taking a picture of a round tortilla with the S22 Ultra, S23 Ultra and the S24 Ultra. The cameras (or processing) DO try to sharpen the features of the tortilla to make it look more "moon-like" but in no scenario could I get any of the cameras to plop in a fake image of the moon. That's not what's going on at all. Same thing with a pure white circle on a black background. Samsung is not inserting fake images of the moon. It is using "AI enhancements" to pull out features when it detects you are taking a moon pic.
I got it to work: ua-cam.com/video/R_xf2TKU7ic/v-deo.html
I've never heard you both discuss your thoughts on Vivian Maier. As you probably know, she was the woman whose (truly amazing) pictures were discovered after she died. She was not famous for her ability until after her death. This brings up a topic that I am constantly thinking about and that is the effect of others seeing your work and how that effects how you take future pictures. I really love your show and I especially like the somewhat existential questions (like the ones discussed in this episode) that you present to the community. Thanks! - Bryant
If these companies start identifying when A.I. filters are used, the 304s on TikTok and Instagram are in trouble.
Photograph=an image created by capturing light. AI images are something else. AI images might replace photographs in many cases, but they are not photographs.
A lot of your discussion was about how much processing you can do of a photograph before disclosing. AI image were never photos at all.
All the discussion about "optical illusion" type photographs and mis-representation are not addressing the Photograph vs AI Image issue at all. Maybe not your goal, but ...
I'm a hobbyist and I've been photographing for several years now but this subject goes beyond photography. IMHO is about perception of reality.
English is my second language so bear with me as this is a very interesting and difficult topic.
Here is what I think: if you or anyone else decide to touch the concept of reality, first and foremost it needs to be defined. So... what's reality and what should be considered as real?
In my opinion, once we press that button in the camera and have an image taken, it is a depiction of reality, therefore, is that considered real or fiction since it is only a snapshot?
Is what we see completely real? Let's not even talk about post image manipulation.
First, is it time a metric that we could use as a reality validation? Is the image considered real or fake due to the fact that all images belong to the past? Is reality intertwined with time? One could take a landscape photo of a sky and the next second could have a cloud. Does that mean when you look at the photo that it is still relevant to reality.... after a few days?
This is a very difficult subject.
Tony and Chelsea.... I ABSOLUTELY love your channel. I think I found you guys since you started.
Thank you so much!
This is an important conversation to have. We, as professional photographers, have a profound responsibility to the truth in a world that increasingly views it with apathy. Another great video, guys!
Glad someone understands the bigger picture here!
It's not often I see hard solipsism being marshalled as a defence of a marketing lie. Kudos, Samsung~?
My second thought is that no version of any image is “real” as no two single people have the exact same quality vision and no one’s mind interprets the raw information from your senses the same way and therefore no two people on the planet ever see the exact same image ever, but that is no excuse for not simple disclosing the fact that the image is “enhanced” as all authentic creators should do, but again this is a corp and not an individual artist. So again, it’s the lie that is offensive and not the message.
I think that Samsung guy got it all backwards, capturing a picture is what makes it real. Your eyes can deceive you, your memory can fail you, but a captured image should be proof that whatever you captured did happen.