Should AI generated images have identifying characteristics?

If you would like to post, you'll need to register. Note that if you have a BCG store account, you'll need a new, separate account here (we keep the two sites separate for security purposes).

Should they be identified?

  • No

    Votes: 1 7.1%
  • Yes

    Votes: 13 92.9%

  • Total voters
    14

sh1209

Well-known member
Supporting Member
Marketplace
With the recent invention of Adobe firefly and several other photography AI generators, should there be a way those images be identified as such? For instance, if I am in the backwoods of Missouri and create a cheetah in the background, should it be mandated these images be identified in some manner? I think it is a good idea because like anything else, it is probably going to be abused. I’m not referring to the remove tool or anything like that and have absolutely no problem with removal of distractions and things of that nature. I think were the gray area really starts whenever you start adding complete subjects to an image that you did not photograph. If there becomes a market for AI generated artwork, then that’s fine. I think Adobe in particular should have some way of baking that information into the metadata of the image, If heavily modified. I have not downloaded the beta version of Photoshop to even try this yet, but I am extremely impressed with the remove tool they have implemented. I just really have no desire to generate things I didn’t photograph. I’m just curious what others think of this?
 
FYI in articles will find statements such as "Adobe says that artwork created using Firefly models will contain metadata indicating that it’s partially — or wholly — AI-generated." I haven't seen examples or an official release on it.
I hope you’re right because I can see this being a problem if something isn’t added for identification.
 
Adobe Annoucement

  • Advocating for open standards: Adobe founded the Content Authenticity Initiative (CAI) to create a global standard for trusted digital content attribution. With more than 900 members worldwide, the role of CAI has never been more important. Adobe is pushing for open industry standards using CAI’s open-source tools that are free and actively developed through the nonprofit Coalition for Content Provenance and Authenticity (C2PA). These goals include a universal “Do Not Train” Content Credentials tag in the image’s Content Credential for creators to request that their content isn’t used to train models. The Content Credentials tag will remain associated with the content wherever it is used, published or stored. In addition, AI generated content will be tagged accordingly.
 
This is a difficult topic as it borders on what I might loosely term freedom of expression. Almost every photo I see these days has had some, however minimal, post processing. Where would the line be drawn?
Does it matter?
I'm not being argumentative, just food for thought.
It certainly matters in my opinion because I could go out and post a rare bird for instance that I didn’t photograph. For general vacation photos and things like that it probably doesn’t matter. If I am being open about posting such an image then there’s no issue. Take photography contests for instance, not all require the raw photo.I think it’s definitely a grey area but as you stated can be used for artistic purposes as well.
 
This is a difficult topic as it borders on what I might loosely term freedom of expression. Almost every photo I see these days has had some, however minimal, post processing. Where would the line be drawn?
Does it matter?
I'm not being argumentative, just food for thought.
I guess I’m old school as I dont add anything to a photo I didn’t photograph. I certainly remove distractions and noise which could also be interpreted as wrong by some people.
 
THere's the Michael Jackson song released after he passed. Several others too....now there 's a new Beatles release due after a so-called AI assisted in getting John Lennon's voice from a "lost" tape. Where will it end?
You are right but there's a million views of every kind.
Personally, I don't edit. My images are rubbish enough without me messing them up further! ;)
 
THere's the Michael Jackson song released after he passed. Several others too....now there 's a new Beatles release due after a so-called AI assisted in getting John Lennon's voice from a "lost" tape. Where will it end?
You are right but there's a million views of every kind.
Personally, I don't edit. My images are rubbish enough without me messing them up further! ;)
I think the majority of us are honest and wouldn’t abuse it but there’s always a few lol
 
The reality is different for each us. I just turned 65. My father played violin in the Academy Awards orchestra pit for 28 shows. When I was 12 we were going to a wedding so it was 1970. He couldn't understand Rock n Roll, or the Beatles who were played for much of the wedding festivities. What my kids think is ok vs what I think differs. The only constant is my life is that our Bengal cats think it's ok to wake us up at 3am.

I have learned to find gratitude and joy in each day; and accept the change as change and hope for the best. At least in the US, we are seeing a huge degradation of integrity. If I let it overtake me, then I will miss out on the good.

Part of my joy is a creative process and learning and why I visit here to see the images and learn. I do applaud Adobe for they balance they are seeking. If I see an image that brings a smile, I am ok if it is part from nature and part from fantasy.
 
If there becomes a market for AI generated artwork, then that’s fine.

Just as a data point, and believing this analysis, the "market for AI" is here and big. Consumers buy images not caring whether they were generated by tinplate or machine learning.

There was a (failed) movement to identify fashion shots altered by Photoshop and I think the fate of identifying AI in pictures will end up the same.

 
The reality is different for each us. I just turned 65. My father played violin in the Academy Awards orchestra pit for 28 shows. When I was 12 we were going to a wedding so it was 1970. He couldn't understand Rock n Roll, or the Beatles who were played for much of the wedding festivities. What my kids think is ok vs what I think differs. The only constant is my life is that our Bengal cats think it's ok to wake us up at 3am.

I have learned to find gratitude and joy in each day; and accept the change as change and hope for the best. At least in the US, we are seeing a huge degradation of integrity. If I let it overtake me, then I will miss out on the good.

Part of my joy is a creative process and learning and why I visit here to see the images and learn. I do applaud Adobe for they balance they are seeking. If I see an image that brings a smile, I am ok if it is part from nature and part from fantasy.
I totally agree with everything you’re saying, and I do occasionally create composites within Photoshop and things like that and enjoy all new technology to be honest. The problem is, just stating a hypothetical here if there is a rare bird in the state I live in they only three have been spotted in the last 50 years And I go and generate an image of that saying I took that photo that’s just plain wrong. Thats the type of scenario I’m referring to, not adding a flower to a bird photo or something like that which is completely harmless. It’s whenever you didn’t even photograph the subject, and or taking credit for it that I have the issue with.
 
I’ve replaced skies in images over the years but have always just pointed my camera in a different direction, snapped the sky, then added it in layers with photoshop. For what it’s worth the new removal tool within Photoshop is honestly one of the best additions I’ve seen in a while that talk and take the place of two or three of the older tools in most situations. I am definitely one that likes new tech and adapts very quickly to it, but I think there needs to be some sort of line between right and wrong
 
Last edited:
This is a great question. I’m torn to the point I can’t vote yes or no. To the extent of adding an animal to a scene that wasn’t there in the original I see as problematic mostly because it will be harder to tell what’s real or not. It kind of diluted the talent of actually taking the picture.

Like you I’m ok with removing sticks and cleaning up a background, I’m ok with say removing a person or distracting element from a travel photo etc.

Where I struggle is adding things. I’ve played with it and while it’s typically not a single push of a button (yet) it can at times do some pretty amazing things.

For me it’s how we look at photography as art or is photography simply skilled documentary.

I think this is where we have to decide are we creating art or not. I struggle with looking at my photography as art but in some ways if we don’t elevate our talents to artistic then it diminishes what we did vs someone snapping a pic with a phone.

It’s easier for me to think a beautiful landscape photo as art and a picture of a bird not being art which isn’t fair but it’s a struggle in my brain.

I’ve been working hard at changing my style of bird photography to make it more artistic than documentary. Having elements in the scene and backing off a bit are some of those changes. So adding in post vs setting it up in the original scene is the question. So far my experience is setting up the scene originally is still delivering better results and feels like a more artistic approach.

For me putting it in the meta data comes down to if we want to know what’s real or not or can we except the final image as art and it doesn’t matter. Not sure about that yet. I guess all of it can and has been done for a long time so what’s changed is it’s easier for people to do now.

So I’ll be interested what everyone says. It’s been a struggle excepting the new tech.
 
Last edited:
This is a difficult topic as it borders on what I might loosely term freedom of expression. Almost every photo I see these days has had some, however minimal, post processing. Where would the line be drawn?
Does it matter?
I'm not being argumentative, just food for thought.
I absolutely think it matters - I mostly avoid the issue by focusing on competitions only now. I get too frustrated with "genuinely made up AI seahorses" getting more attention than actual great photographs of seahorses. I avoid that by not shooting for joe public, but for conservation and competition. AI can't be 100% dismissed in competitions, but they can put ground rules in to minimise it.

I think the delineation is very very clear:
Photo adjustments vs photo manipulation:
* Photo adjustments: you make adjustments to the photo itself, only to the photo taken between the lens and the camera sensor
* Photo manipulation: you add/remove artefacts to the photo that exist outside of the lens and sensor
* Photo make-believe: just magic something up with an AI prompt

I'm seeing this massive trend of people insisting on calling it "a photo" when in reality, it is photographic manipulation / composites, aka. "a composite".

I don't think anyone cares if it is a photo or a composite, but please just be genuine in your approach to both social media and your photography :)

(sorry if this felt like a rant, but the number of people who get conned into thinking a composite is a photo, or an ai generated image is a photo, is too high! :D)
 
Last edited:
This is a great question. I’m torn to the point I can’t vote yes or no. To the extent of adding an animal to a scene that wasn’t there in the original I see as problematic mostly because it will be harder to tell what’s real or not. It kind of diluted the talent of actually taking the picture.

Like you I’m ok with removing sticks and cleaning up a background, I’m ok with say removing a person or distracting element from a travel photo etc.

Where I struggle is adding things. I’ve played with it and while it’s typically not a single push of a button (yet) it can at times do some pretty amazing things.

For me it’s how we look at photography as art or is photography simply skilled documentary.

I think this is where we have to decide are we creating art or not. I struggle with looking at my photography as art but in some ways if we don’t elevate our talents to artistic then it diminishes what we did vs someone snapping a pic with a phone.

It’s easier for me to think a beautiful landscape photo as art and a picture of a bird not being art which isn’t fair but it’s a struggle in my brain.

I’ve been working hard at changing my style of bird photography to make it more artistic than documentary. Having elements in the scene and backing off a bit are some of those changes. So adding in post vs setting it up in the original scene is the question. So far my experience is setting up the scene originally is still delivering better results and feels like a more artistic approach.

For me putting it in the meta data comes down to if we want to know what’s real or not or can we except the final image as art and it doesn’t matter. Not sure about that yet. I guess all of it can and has been done for a long time so what’s changed is it’s easier for people to do now.

So I’ll be interested what everyone says. It’s been a struggle excepting the new tech.
I agree with everything you were saying, and I do like using such tools, but if I present an image, I want to be upfront about it and be honest about it no problem I have is when people start portraying subjects they didn’t photograph and saying they did that’s where it is totally absolutely wrong in my opinion.
 
I absolutely think it matters - I mostly avoid the issue by focusing on competitions only now. I get too frustrated with "genuinely made up AI seahorses" getting more attention than actual great photographs of seahorses. I avoid that by not shooting for joe public, but for conservation and competition. AI can't be 100% dismissed in competitions, but they can put ground rules in to minimise it.

I think the delineation is very very clear:
Photo adjustments vs photo manipulation:
* Photo adjustments: you make adjustments to the photo itself, only to the photo taken between the lens and the camera sensor
* Photo manipulation: you add/remove artefacts to the photo that exist outside of the lens and sensor
* Photo make-believe: just magic something up with an AI prompt

I'm seeing this massive trend of people insisting on calling it "a photo" when in reality, it is photographic manipulation / composites, aka. "a composite".

I don't think anyone cares if it is a photo or a composite, but please just be genuine in your approach to both social media and your photography :)

(sorry if this felt like a rant, but the number of people who get conned into thinking a composite is a photo, or an ai generated image is a photo, is too high! :D)
That’s pretty much where I stand as well. I do remove noise and an occasional distraction but typically don’t go beyond that.
 
I agree with everything you were saying, and I do like using such tools, but if I present an image, I want to be upfront about it and be honest about it no problem I have is when people start portraying subjects they didn’t photograph and saying they did that’s where it is totally absolutely wrong in my opinion.
Yeah if lying about it then that’s wrong. I do wonder though as I struggle with this question is does an artist have to outline their process for the finished product? I think the answer is no so why should a photographer who is also an artist? Not being argumentative just can see how some struggle with this.
 
I’m strongly in favor of irrevocably noting in the metadata whether an image has been manipulated by an AI program. People can then decide for themselves whether, or how, such images should be used and appreciated. For me, disclosure is the key; then, we can all proceed according to our own ethics and preferences. I have no interest in completely fake images or overly manipulated images, but there is no way to define “overly manipulated.” It’s just a personal preference. I will definitely use Adobe’s noise reduction and content removal AI tools, but I’m not going to add things that weren’t actually in the image (outside of stretching the canvas a bit; see, strict definitions are hard).

I don’t really see how any of this this intrudes upon freedom of expression, at least constitutionally speaking. The constitution only applies to and prohibits governments from taking certain actions, like censorship or prosecuting someone without due process. But the constitution has no application in the private world.
 
I do believe that AI-generated images, or those that have had things added or removed should have some identifying signal as such. Adobe's Content Authenticity Initiative is meant to address this, but the initiative is way behind the curve right now.


The challenge with the discussions around AI generated content is that each party is coming to the table with different needs and expectations for what a photo is, or what a photo illustration should be. News photographers are aghast at the idea of altering an image beyond toning adjustments, while astrophotographers wouldn't be where they're at today without significant image stacking and alterations. Branding and marketing have been manipulating photos for so long it's expected at this point.


In the absence of CAI, authenticity will come down to individual photographers publicizing what manipulation (if any) they've done to a photo.
 
I think it should be part of one's own ethical standards. It should certainly be part of a publications journalistic standards. But I think current law probably mostly covers situations where altered images are used for fraud, forgery, libel, plaigerism etc. I don't use Ai, Or do I? I use lightroom denoise, isn't that AI? I use Photoshop Remove. Isn't that AI. If I extend my own canvas with generative fill, isn't that AI. Hard to draw the line, except I have my ethics and I stick with them.
 
I’m strongly in favor of irrevocably noting in the metadata whether an image has been manipulated by an AI program. People can then decide for themselves whether, or how, such images should be used and appreciated. For me, disclosure is the key; then, we can all proceed according to our own ethics and preferences. I have no interest in completely fake images or overly manipulated images, but there is no way to define “overly manipulated.” It’s just a personal preference. I will definitely use Adobe’s noise reduction and content removal AI tools, but I’m not going to add things that weren’t actually in the image (outside of stretching the canvas a bit; see, strict definitions are hard).

I don’t really see how any of this this intrudes upon freedom of expression, at least constitutionally speaking. The constitution only applies to and prohibits governments from taking certain actions, like censorship or prosecuting someone without due process. But the constitution has no application in the private world.
It seems it would be a fairly simple process for adobe to enable such a feature and I’m definitely in the same camp as you agreeing it should somehow be disclosed.
 
I think it just depends on the setting. I do not make a note on Instagram that I combined a couple landscape images to create one photo that contains detail in the lights and darks. But I would disclose that in a competition and to anyone who asked.

I somehow managed to mess this post up. It was a response to @DavidT
 
I think it just depends on the setting. I do not make a note on Instagram that I combined a couple landscape images to create one photo that contains detail in the lights and darks. But I would disclose that in a competition and to anyone who asked.

I somehow managed to mess this post up. It was a response to @DavidT
I could definitely see competitors requiring disclosure and require a raw image and I think some already do. I wonder if you shoot in jpeg what they do?

Instagram is a beast of its own as when you go to upload a photo they provide you will all kinds of options to make the picture not what you took. They are endorsing the practice and many over saturate, and use filters that totally alter the image. It is know as an IG style.
 
A very stretchable subject. Metadata included in an image doesn't help if the image goes viral on FB or Instagram and no one can see the Metadata at a glance.

I think for us as photographers it means something different than it does to graphic designers / digital artists. I still like shooting my own sjit. But I also enjoy digital art - not to be confused with my photography and for that I simply add Ai to my signature. It's so warped no one would mistake it for reality anyway.
 
Back
Top