Show Posts
1
Adobe Stock / Adobe Rejections for Possibly not declaring Generative AI
« on: January 19, 2024, 14:49 »
Today I received rejection notices from Adobe Stock that stated the following...
Possible reasons:
- Non compliant use of another artists name.
- Undeclared Generative AI Content.
- Content not compliant with overall guidelines
I think I understand what the first two reasons mean. The third one seems pretty nebulous and vague. Could the third reason have nothing to do with generative AI whatsoever and that the photo was rejected for more traditional reasons such as image quality, noise, or whatever? I don't know. Now, regarding the second possible reason above, Adobe stated in a document called, "10 tips for getting your generative AI images approved for sale on Adobe Stock," the following in tip #5...
If you are using generative fill on a photograph only to extend a background or remove distracting objects, people, or intellectual property, then you do not need to label the image as generative AI.
The document can be found here... https://stock.adobe.com/pages/artisthub/pdf/generative-ai-guide.pdf
I shot and processed the photographs in question. I did use generative AI fill tools very sparingly to 1) Add a little non-descript edge content (grass, sky, etc.) where I needed to straighten images via rotation and ended up having some small blank areas to fill in the corners, and 2) remove some small distractions such as sensor spots, bright pixels, or a power line or two going through tree branches, etc.
My intentions when editing are certainly not to deceive or to "pull a fast one." Not so long ago I would have just used spot healing and content aware fill for such purposes (and still do). I used generative AI tools on these photos because I thought they might yield better results than my typical way of dealing with fills and spot healing, and because, after reading the Adobe-published doc cited above, I thought it was perfectly permissible and not a violation for this type of editing. So, I am confused.
I know the rejections reason was not for reason #1 (unless a robot decided that "Cape Hatteras" in the key words was a person's name or some other such robot silliness). I am not sure if the rejections were due to reasons #2 and/or #3 above. If #3 (provided it has nothing to do with generative AI), it is no big deal. I've had a few rejections regarding quality problems now and then. Life goes on. But, if the rejections are due to reason #2, well, I'll stop using generative AI altogether and go back to using only content aware fill when an image calls for that type of correction. I certainly do not want to be banned from submitting to Adobe Stock over something like this and am having trouble reconciling the rejection notice with what is stated in the Adobe doc that clearly states that this type of generative AI usage is permissible. Is some minor generative AI healing or edge fill allowed or not?
Again, I want to follow the rules and not be banned from submitting to Adobe Stock. But, what exactly are the rules?
Any thoughts or comments on similar experiences are greatly appreciated. Thanks!
Possible reasons:
- Non compliant use of another artists name.
- Undeclared Generative AI Content.
- Content not compliant with overall guidelines
I think I understand what the first two reasons mean. The third one seems pretty nebulous and vague. Could the third reason have nothing to do with generative AI whatsoever and that the photo was rejected for more traditional reasons such as image quality, noise, or whatever? I don't know. Now, regarding the second possible reason above, Adobe stated in a document called, "10 tips for getting your generative AI images approved for sale on Adobe Stock," the following in tip #5...
If you are using generative fill on a photograph only to extend a background or remove distracting objects, people, or intellectual property, then you do not need to label the image as generative AI.
The document can be found here... https://stock.adobe.com/pages/artisthub/pdf/generative-ai-guide.pdf
I shot and processed the photographs in question. I did use generative AI fill tools very sparingly to 1) Add a little non-descript edge content (grass, sky, etc.) where I needed to straighten images via rotation and ended up having some small blank areas to fill in the corners, and 2) remove some small distractions such as sensor spots, bright pixels, or a power line or two going through tree branches, etc.
My intentions when editing are certainly not to deceive or to "pull a fast one." Not so long ago I would have just used spot healing and content aware fill for such purposes (and still do). I used generative AI tools on these photos because I thought they might yield better results than my typical way of dealing with fills and spot healing, and because, after reading the Adobe-published doc cited above, I thought it was perfectly permissible and not a violation for this type of editing. So, I am confused.
I know the rejections reason was not for reason #1 (unless a robot decided that "Cape Hatteras" in the key words was a person's name or some other such robot silliness). I am not sure if the rejections were due to reasons #2 and/or #3 above. If #3 (provided it has nothing to do with generative AI), it is no big deal. I've had a few rejections regarding quality problems now and then. Life goes on. But, if the rejections are due to reason #2, well, I'll stop using generative AI altogether and go back to using only content aware fill when an image calls for that type of correction. I certainly do not want to be banned from submitting to Adobe Stock over something like this and am having trouble reconciling the rejection notice with what is stated in the Adobe doc that clearly states that this type of generative AI usage is permissible. Is some minor generative AI healing or edge fill allowed or not?
Again, I want to follow the rules and not be banned from submitting to Adobe Stock. But, what exactly are the rules?
Any thoughts or comments on similar experiences are greatly appreciated. Thanks!
